{"schema_version":"onlylabs.public_analysis.v1","url":"https://onlylabs.fyi/analysis/friendliai","json_url":"https://onlylabs.fyi/analysis/friendliai/analysis.json","evidence_json_url":"https://onlylabs.fyi/analysis/friendliai/evidence.json","generated_at":"2026-06-27T22:34:22.831Z","analysis":{"org_slug":"friendliai","url":"https://onlylabs.fyi/analysis/friendliai","json_url":"https://onlylabs.fyi/analysis/friendliai/analysis.json","evidence_json_url":"https://onlylabs.fyi/analysis/friendliai/evidence.json","dossier_url":"https://onlylabs.fyi/labs/friendliai","org":{"slug":"friendliai","name":"FriendliAI","category":"neocloud","category_label":"Neocloud","homepage_url":"https://friendli.ai"},"title":"FriendliAI analysis","summary":"FriendliAI is an AI inference infrastructure company entering an aggressive commercialization phase, signaled by a $20M funding round, a rapid SDK iteration cadence with breaking API changes across all serving tiers, the launch of a public OpenAPI schema, and day-zero support for frontier open-weight models. The dual-hub (Seoul/San Francisco) hiring pattern reveals simultaneous investment in core inference engine…","markdown":"## Thesis\n\nFriendliAI is an AI inference infrastructure company entering an aggressive commercialization phase, signaled by a $20M funding round [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735), a rapid SDK iteration cadence with breaking API changes across all serving tiers [P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2), the launch of a public OpenAPI schema [P4](https://github.com/friendliai/friendli-openapi), and day-zero support for frontier open-weight models [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra)[W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz)[W3](https://friendli.ai/blog/kimi-k2-6)[W4](https://friendli.ai/blog/deepseek-v4-pro-flash)[W5](https://friendli.ai/blog/GLM-5.2). The dual-hub (Seoul/San Francisco) hiring pattern [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1) reveals simultaneous investment in core inference engine performance—GPU kernel engineering, compiler infrastructure, cross-vendor (NVIDIA/AMD) parity [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735)—and a GTM buildout for enterprise sales in the US [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1)[P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706)[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da). The fork portfolio maps a full-stack inference validation chain: GPU kernel libraries (CUTLASS, FlashInfer, TensorRT-LLM) [E27](https://github.com/friendliai/TensorRT-LLM)[E54](https://github.com/friendliai/flashinfer)[E55](https://github.com/friendliai/cutlass), evaluation harnesses (EleutherAI, OpenAI simple-evals) [P8](https://github.com/friendliai/lm-evaluation-harness)[E59](https://github.com/friendliai/simple-evals-archived), and agent/application frameworks (LangChain, LlamaIndex) [P9](https://github.com/friendliai/langchain)[P10](https://github.com/friendliai/llama_index). Together, these signals position FriendliAI as a \"deploy and forget\" inference platform competing on performance and developer experience for the open-weight model ecosystem.\n\n## Signal desks\n\n### Hiring\n\n- **Inference Systems (GPU Kernel & Inference Engine):** Active hiring in both Seoul and San Francisco for engineers to design custom GPU kernels (CUDA/ROCm/HIP), build kernel compilers, memory planners, and runtimes targeting transformer/diffusion workloads across NVIDIA and AMD hardware [P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735). Roles cite 450k+ Hugging Face model support, \"inventors of continuous batching,\" and FP8/FP4 quantized kernel work as differentiators [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735). This signals deep investment in inference speed as the core technical moat.\n- **Core Product (Backend, Full-Stack, QA):** Hiring Senior Backend Engineers in Seoul and San Francisco for multi-tenant SaaS with authentication, RBAC, billing, and multi-cloud orchestration atop PostgreSQL and ClickHouse [P17](https://jobs.ashbyhq.com/friendliai/5fd918c3-70c8-44f1-b080-2aef9520b312)[P24](https://jobs.ashbyhq.com/friendliai/d5288f06-df73-407d-9d32-3eb96d735a51). Full-Stack (React/Next.js, TypeScript, FastAPI) and QA (pytest, Locust, Playwright, LLM-specific inference quality validation) roles round out the platform buildout [P21](https://jobs.ashbyhq.com/friendliai/25429014-12bd-4665-a435-367d71a785c2)[P19](https://jobs.ashbyhq.com/friendliai/41859c5f-fd60-40a3-9966-39a78852008b). Implies an enterprise-grade SaaS control plane maturing rapidly.\n- **Applied Engineering (AI Agents, Python Developer Tools):** Hiring AI Agent engineers in both Seoul and San Francisco to build a Friendli Agent API, document understanding, advanced RAG, and reference agent applications [P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362). Python Developer Tools engineers are tasked with SDK/CLI ownership, PyPI distribution, and monorepo management [P23](https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737). Signals productization of agentic workflows as a platform layer atop the inference engine.\n- **Solutions & GTM (Account Executive, Solutions Architects, Developer Advocate, Customer Success):** San Francisco-based Account Executive for enterprise inference deals [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1). Two Solutions Architect roles—one for inference/Kubernetes deployment, one for model integration/agentic frameworks [P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706)[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da). Developer Advocates in both Seoul and San Francisco for community events, hackathons, and content [P28](https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09)[E9](https://jobs.ashbyhq.com/friendliai/8d4ac792-7b6b-4503-84a0-a321ee68fe71)[E11](https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09). Contract Customer Success Engineer in Seoul [P25](https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7). Senior Product Manager in Seoul to own roadmap across model APIs, deployment workflows, and developer-facing features [P12](https://jobs.ashbyhq.com/friendliai/a6364382-e30a-4a30-919d-2c6b6c78185a). This is a full GTM buildout targeting US enterprise adoption.\n\n### Forks\n\n- **GPU Kernel & Inference Libraries:** NVIDIA/TensorRT-LLM [E27](https://github.com/friendliai/TensorRT-LLM), NVIDIA/cutlass [E55](https://github.com/friendliai/cutlass), flashinfer-ai/flashinfer [E54](https://github.com/friendliai/flashinfer). These map directly to the kernel optimization and compiler work described in Inference Systems job postings [P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735).\n- **Evaluation Harnesses:** EleutherAI/lm-evaluation-harness (forked with custom \"periflow\" model integration supporting sequential and async evaluation) [P8](https://github.com/friendliai/lm-evaluation-harness), openai/simple-evals [E59](https://github.com/friendliai/simple-evals-archived). The lm-evaluation-harness fork includes a PeriFlow-specific task table and custom request URL configuration, indicating internal eval infrastructure built on community standards [P8](https://github.com/friendliai/lm-evaluation-harness).\n- **Agent & Application Frameworks:** langchain-ai/langchain [P9](https://github.com/friendliai/langchain), run-llama/llama_index [P10](https://github.com/friendliai/llama_index), ShishirPatil/gorilla [E58](https://github.com/friendliai/gorilla), weaviate/weaviate-io [E60](https://github.com/friendliai/weaviate-io). These support the AI Agents hiring thesis—FriendliAI is building integrations and reference architectures on top of the dominant agent frameworks [P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362).\n- **Infrastructure & Observability:** ContainerSolutions/locust_exporter (Prometheus metrics for Locust load testing) [P11](https://github.com/friendliai/locust_exporter), ScalingIntelligence/tokasaurus [E57](https://github.com/friendliai/tokasaurus), earendil-works/pi [E42](https://github.com/friendliai/pi-mono). The Locust exporter fork aligns with the QA role's explicit use of Locust for scalability testing [P19](https://jobs.ashbyhq.com/friendliai/41859c5f-fd60-40a3-9966-39a78852008b).\n- **Content & Ecosystem:** huggingface/blog [E28](https://github.com/friendliai/huggingface-blog), anomalyco/opencode [E44](https://github.com/friendliai/opencode), anomalyco/models.dev [E46](https://github.com/friendliai/models.dev). Low-signal forks, possibly for blog infrastructure or community tooling exploration.\n\n### Releases\n\n- **friendli-python SDK (v0.10.5 through v0.13.2):** Rapid iteration from June 2025 through June 2026, with Speakeasy CL auto-generation from an OpenAPI spec [P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0)[P2](https://github.com/friendliai/friendli-python/releases/tag/v0.13.1)[P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2)[P5](https://github.com/friendliai/friendli-python/releases/tag/v0.12.6)[P6](https://github.com/friendliai/friendli-python/releases/tag/v0.12.8). v0.13.0 removed `serverless.knowledge.retrieve()` and `serverless.model.list()` as breaking changes [P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0). v0.13.2 introduced breaking changes to `stream` parameters across all serving tiers—container, dedicated, and serverless—for chat, completions, and audio endpoints [P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2). This signals an aggressive API consolidation wave across the entire product surface.\n- **friendli-openapi (new repo, June 2026):** Public OpenAPI schema repository created 2026-06-22 [P4](https://github.com/friendliai/friendli-openapi)[E4](https://github.com/friendliai/friendli-openapi)[E5](https://github.com/friendliai/friendli-openapi). Provides a machine-readable spec for the Friendli API, enabling code generation and third-party integration [P4](https://github.com/friendliai/friendli-openapi). This is a developer-experience investment that supports the SDK automation pipeline and external tooling ecosystem.\n- **Friendli Model APIs (documented June 2026):** Serverless inference interface positioned as an OpenAI-compatible rental service that abstracts GPU infrastructure, described as \"no capacity planning, no GPU orchestration, no cold starts\" [W5](https://friendli.ai/blog/GLM-5.2)[W6](https://friendli.ai/docs/guides/model-apis/introduction).\n\n### Talking\n\n- **Day-Zero Model Support as Market Positioning:** FriendliAI's blog and LinkedIn content consistently emphasizes \"Day 0\" or immediate availability for frontier open-weight models—Nemotron 3 Ultra (550B) [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra), MiniMax-M3 (1M-token context, multimodal) [W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz), Kimi K2.6 [W3](https://friendli.ai/blog/kimi-k2-6), DeepSeek V4 Pro/Flash [W4](https://friendli.ai/blog/deepseek-v4-pro-flash), and GLM-5.2 [W5](https://friendli.ai/blog/GLM-5.2). This framing positions FriendliAI as the fastest path from model release to production deployment.\n- **Hugging Face Integration:** Multiple posts highlight one-click deployment from Hugging Face Hub to Friendli Dedicated Endpoints, leveraging a direct Hugging Face partnership [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra)[W3](https://friendli.ai/blog/kimi-k2-6)[W4](https://friendli.ai/blog/deepseek-v4-pro-flash). This is a distribution strategy that reduces customer onboarding friction and taps the open-weight model community.\n- **San Francisco Expansion:** The DeepSeek V4 blog explicitly announces FriendliAI's expansion to San Francisco to \"scale frontier AI inference for open-weight and custom models\" [W4](https://friendli.ai/blog/deepseek-v4-pro-flash), consistent with the heavy SF-based hiring across engineering, sales, and solutions roles [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1)[P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706)[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da)[P19](https://jobs.ashbyhq.com/friendliai/41859c5f-fd60-40a3-9966-39a78852008b)[P20](https://jobs.ashbyhq.com/friendliai/372dbe1e-d6e6-4726-ae09-673ec8916d05)[P21](https://jobs.ashbyhq.com/friendliai/25429014-12bd-4665-a435-367d71a785c2)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362)[P23](https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737)[P24](https://jobs.ashbyhq.com/friendliai/d5288f06-df73-407d-9d32-3eb96d735a51).\n- **Performance and Cost Narrative:** Blog posts frame FriendliAI's value proposition as \"unmatched speed, cost efficiency, and operational simplicity\" [W5](https://friendli.ai/blog/GLM-5.2), with specific technical claims about MiniMax-M3's sparse attention delivering ~9× faster prefill and ~15× faster decode at long context [W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz). No cited third-party benchmarks are provided in this evidence pack to independently verify these claims.\n\n## Shipping\n\nFriendliAI's shipping activity is concentrated in three areas in the evidence window:\n\n1. **Python SDK (friendli-core on PyPI):** A dense release cadence—v0.10.5 (Jun 2025) [E56](https://github.com/friendliai/friendli-python/releases/tag/v0.10.5), v0.10.6 (Jul 2025) [E53](https://github.com/friendliai/friendli-python/releases/tag/v0.10.6), v0.10.7 (Aug 2025) [E52](https://github.com/friendliai/friendli-python/releases/tag/v0.10.7), v0.10.9 (Aug 2025) [E50](https://github.com/friendliai/friendli-python/releases/tag/v0.10.9), v0.11.0 (Sep 2025) [E49](https://github.com/friendliai/friendli-python/releases/tag/v0.11.0), v0.12.1–v0.12.4 (Dec 2025–Jan 2026) [E48](https://github.com/friendliai/friendli-python/releases/tag/v0.12.1)[E47](https://github.com/friendliai/friendli-python/releases/tag/v0.12.2)[E45](https://github.com/friendliai/friendli-python/releases/tag/v0.12.3)[E43](https://github.com/friendliai/friendli-python/releases/tag/v0.12.4), v0.12.6–v0.12.8 (Jun 2026) [E7](https://github.com/friendliai/friendli-python/releases/tag/v0.12.6)[E6](https://github.com/friendliai/friendli-python/releases/tag/v0.12.8), and v0.13.0–v0.13.2 (Jun 2026) [E3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0)[E2](https://github.com/friendliai/friendli-python/releases/tag/v0.13.1)[E1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2). The v0.13.x series introduced breaking changes across all serving tiers, indicating a deliberate API rationalization [P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0)[P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2). All releases are Speakeasy-generated from an OpenAPI spec, suggesting a design-first API development workflow [P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0)[P2](https://github.com/friendliai/friendli-python/releases/tag/v0.13.1)[P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2).\n\n2. **OpenAPI Specification (friendli-openapi):** A new standalone repository published June 2026 with an Apache-2.0 license, exposing the Friendli API schema for external code generation and tooling integration [P4](https://github.com/friendliai/friendli-openapi)[E4](https://github.com/friendliai/friendli-openapi)[E5](https://github.com/friendliai/friendli-openapi).\n\n3. **Serverless Product (Friendli Model APIs):** Documented in June 2026 as an OpenAI-compatible, serverless inference interface abstracting GPU management [W5](https://friendli.ai/blog/GLM-5.2)[W6](https://friendli.ai/docs/guides/model-apis/introduction).\n\nOther visible repos—friendli-client (deprecated) [E19](https://github.com/friendliai/friendli-client), periflow-cli [E23](https://github.com/friendliai/periflow-cli), FAI-Model [E18](https://github.com/friendliai/FAI-Model), friendli-model-optimizer [E22](https://github.com/friendliai/friendli-model-optimizer), LLMServingPerfEvaluator [E20](https://github.com/friendliai/LLMServingPerfEvaluator), aipm [E21](https://github.com/friendliai/aipm), friendli-gradio [E51](https://github.com/friendliai/friendli-gradio), examples [E25](https://github.com/friendliai/examples), and llm-hackathon-tutorial [E26](https://github.com/friendliai/llm-hackathon-tutorial)—have not had recent release activity in this evidence pack, though they collectively demonstrate a history of tooling, benchmarking, and community engagement investment.\n\n## Research themes\n\nEvidence suggests FriendliAI's research investment focuses on inference systems engineering rather than model training:\n\n- **GPU Kernel Optimization:** Job descriptions and fork activity converge on custom kernel development for GEMM, attention, and routing operations, including FP8/FP4 reduced-precision kernels and cross-vendor (NVIDIA/AMD) parity work [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735)[E27](https://github.com/friendliai/TensorRT-LLM)[E54](https://github.com/friendliai/flashinfer)[E55](https://github.com/friendliai/cutlass).\n- **Kernel Compiler & Runtime:** Multiple Inference Engine roles reference work on a proprietary kernel compiler, memory planner, and runtime—consistent with building compiler-like optimization layers rather than hand-tuning individual kernels [P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe).\n- **Continuous Batching & Dynamic Shape Compilation:** Job posts claim FriendliAI employs the \"inventors of continuous batching\" and list dynamic shape compilation and memory planning as preferred experience [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735)[P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe).\n- **Agentic AI Systems:** Applied Engineering roles building agent APIs, document understanding, and advanced RAG [P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362) suggest applied research into production-grade agent orchestration atop the inference layer.\n- **Eval Infrastructure:** The lm-evaluation-harness fork with custom PeriFlow async evaluation support [P8](https://github.com/friendliai/lm-evaluation-harness) and the LLMServingPerfEvaluator repo [E20](https://github.com/friendliai/LLMServingPerfEvaluator) indicate sustained investment in benchmarking and performance validation tooling.\n\nNo cited evidence in this pack indicates FriendliAI is training foundation models.\n\n## Hiring & scaling\n\nFriendliAI is hiring across 17+ distinct roles spanning six teams: Inference Systems, Core Product, Applied Engineering, Solutions Architect, Sales, Marketing, Product, and Infrastructure [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1). Key patterns:\n\n- **Dual-hub geography:** Seoul (OnSite) and San Francisco (Hybrid/Remote). Engineering roles appear in both locations, but GTM roles (Account Executive, both Solutions Architects) are exclusively San Francisco-based [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1)[P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706)[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da), while Product leadership (Senior PM) and customer-facing contract roles sit in Seoul [P12](https://jobs.ashbyhq.com/friendliai/a6364382-e30a-4a30-919d-2c6b6c78185a)[P25](https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7). This maps to an \"engine in Seoul, sell in SF\" structure.\n- **Inference Systems is the most repeated hiring theme:** Three distinct role types (Inference Engine, GPU Kernel) posted in both Seoul and SF [P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735), with some listings explicitly citing $20M in funding as context for scaling [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735).\n- **Applied Engineering as a distinct team:** AI Agents and Python Developer Tools roles signal a product-engineering function separate from Core Product, focused on developer experience, SDKs, and agentic application layer [P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362)[P23](https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737)[P25](https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7).\n- **Enterprise GTM buildout:** Account Executive with \"uncapped commission structure tied to enterprise deal value\" and equity [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1), plus two Solutions Architect specializations (infrastructure/deployment and model integration/agents) [P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706)[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da), and Developer Advocates in both hubs [P28](https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09)[E9](https://jobs.ashbyhq.com/friendliai/8d4ac792-7b6b-4503-84a0-a321ee68fe71)[E11](https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09) indicate a full enterprise sales and community funnel.\n- **Contract-to-hire experimentation:** A 6-month contract Customer Success Engineer role in Seoul [P25](https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7) suggests cautious scaling on the customer-facing side before committing to permanent headcount.\n\n## Category implications\n\n- **Infrastructure strategy:** FriendliAI's simultaneous investment in GPU kernel optimization (CUTLASS, FlashInfer, TensorRT-LLM forks) [E27](https://github.com/friendliai/TensorRT-LLM)[E54](https://github.com/friendliai/flashinfer)[E55](https://github.com/friendliai/cutlass), a proprietary kernel compiler [P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b)[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe), and cross-vendor NVIDIA/AMD parity [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735) positions the company as an inference performance competitor to purpose-built serving engines. The \"Friendli Container\" product allowing deployment in private clouds or on-premises via EKS add-on [P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706) suggests a hybrid deployment strategy targeting enterprises that cannot use shared SaaS inference.\n- **Developer experience as product moat:** The release of a public OpenAPI spec [P4](https://github.com/friendliai/friendli-openapi), Speakeasy-generated SDK automation [P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0)[P2](https://github.com/friendliai/friendli-python/releases/tag/v0.13.1)[P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2), dedicated Python Developer Tools hiring [P23](https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737), and one-click Hugging Face Hub deployment [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra)[W3](https://friendli.ai/blog/kimi-k2-6)[W4](https://friendli.ai/blog/deepseek-v4-pro-flash) indicate that API ergonomics and ecosystem integration are being built as competitive differentiators, not afterthoughts.\n- **Agentic platform layer:** AI Agents hiring across two locations [P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6)[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362), the Friendli Agent API, and forks of LangChain and LlamaIndex [P9](https://github.com/friendliai/langchain)[P10](https://github.com/friendliai/llama_index) signal a product thesis that inference is the substrate for an agent platform. This could compete with agent-native platforms if FriendliAI builds proprietary orchestration, memory management, and tool-use primitives into its API.\n- **Open-weight model ecosystem dependency:** All cited blog content promotes day-zero support for third-party open-weight models (Nemotron, MiniMax, Kimi, DeepSeek, GLM) [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra)[W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz)[W3](https://friendli.ai/blog/kimi-k2-6)[W4](https://friendli.ai/blog/deepseek-v4-pro-flash)[W5](https://friendli.ai/blog/GLM-5.2). FriendliAI's value proposition is tied to the continued release and enterprise adoption of frontier open-weight models. A shift toward closed-model dominance or self-hosting by model providers would directly impact this strategy.\n- **Hiring implications for competitors:** The depth of GPU kernel and inference engine hiring—requiring CUDA, ROCm/HIP, and compiler expertise [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735)—suggests that inference optimization talent is a binding constraint. The \"inventors of continuous batching\" claim [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735) indicates this team may include key contributors from the original Orca/continuous batching research lineage.\n\n## Traction highlights\n\n- **$20M funding round** cited in GPU Kernel and Inference Engine job postings as context for team scaling [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735).\n- **450,000+ Hugging Face model support** claimed across job posts [P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b)[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735).\n- **Day-zero deployment** for five major frontier open-weight models in a compressed window: DeepSeek V4 Pro/Flash (May 2026) [W4](https://friendli.ai/blog/deepseek-v4-pro-flash), Kimi K2.6 (May 2026) [W3](https://friendli.ai/blog/kimi-k2-6), Nemotron 3 Ultra (Jun 2026) [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra), GLM-5.2 (Jun 2026) [W5](https://friendli.ai/blog/GLM-5.2), MiniMax-M3 (Jun 2026) [W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz).\n- **Hugging Face partnership** enabling one-click \"Friendli Endpoints\" deployment from the Hub [W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra)[W3](https://friendli.ai/blog/kimi-k2-6)[W4](https://friendli.ai/blog/deepseek-v4-pro-flash).\n- **GitHub community signals:** flagship repos FAI-Model (89 stars) [E18](https://github.com/friendliai/FAI-Model), friendli-client (50 stars, deprecated) [E19](https://github.com/friendliai/friendli-client), LLMServingPerfEvaluator (48 stars) [E20](https://github.com/friendliai/LLMServingPerfEvaluator), and aipm (21 stars) [E21](https://github.com/friendliai/aipm) show modest but non-trivial community interest. The newer friendli-python SDK has 1 star [E29](https://github.com/friendliai/friendli-python), consistent with its recent transition from a Speakeasy-generated internal tool to a public artifact.\n- **17+ open roles** across 6 teams in 2 global hubs as of March–June 2026 [P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1), indicating rapid organizational scaling.\n\nEvidence is thin on: revenue, paying customer counts, inference volume metrics, and independent third-party benchmark comparisons. The job descriptions and blog content provide FriendliAI's own framing of its performance advantages, but no cited external validation (e.g., published benchmark results, customer case studies with named logos, or analyst evaluations) is present in this pack.\n\n## Sources\n\n[P1](https://github.com/friendliai/friendli-python/releases/tag/v0.13.0) friendli-python v0.13.0 release notes, 2026-06-22\n[P2](https://github.com/friendliai/friendli-python/releases/tag/v0.13.1) friendli-python v0.13.1 release notes, 2026-06-22\n[P3](https://github.com/friendliai/friendli-python/releases/tag/v0.13.2) friendli-python v0.13.2 release notes, 2026-06-22\n[P4](https://github.com/friendliai/friendli-openapi) friendli-openapi repository, 2026-06-22\n[P5](https://github.com/friendliai/friendli-python/releases/tag/v0.12.6) friendli-python v0.12.6 release notes, 2026-06-20\n[P6](https://github.com/friendliai/friendli-python/releases/tag/v0.12.8) friendli-python v0.12.8 release notes, 2026-06-20\n[P7](https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1) Account Executive job posting, San Francisco, 2026-06-10\n[P8](https://github.com/friendliai/lm-evaluation-harness) friendliai/lm-evaluation-harness fork, created 2023-08-18\n[P9](https://github.com/friendliai/langchain) friendliai/langchain fork, created 2024-02-15\n[P10](https://github.com/friendliai/llama_index) friendliai/llama_index fork, created 2024-02-23\n[P11](https://github.com/friendliai/locust_exporter) friendliai/locust_exporter fork, created 2024-03-18\n[P12](https://jobs.ashbyhq.com/friendliai/a6364382-e30a-4a30-919d-2c6b6c78185a) Senior Product Manager job posting, Seoul, 2026-03-15\n[P13](https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b) Software Engineer – AI Inference Engine job posting, San Francisco, 2026-03-15\n[P14](https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b) Software Engineer – GPU Kernel job posting, San Francisco, 2026-03-15\n[P15](https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706) Solutions Architect – AI Inference Specialist job posting, San Francisco, 2026-03-15\n[P16](https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da) Solutions Architect – AI Model Specialist job posting, San Francisco, 2026-03-15\n[P17](https://jobs.ashbyhq.com/friendliai/5fd918c3-70c8-44f1-b080-2aef9520b312) Software Engineer – Senior Backend job posting, Seoul, 2026-03-15\n[P18](https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6) Software Engineer – AI Agents job posting, Seoul, 2026-03-24\n[P19](https://jobs.ashbyhq.com/friendliai/41859c5f-fd60-40a3-9966-39a78852008b) QA Engineer job posting, San Francisco, 2026-03-27\n[P20](https://jobs.ashbyhq.com/friendliai/372dbe1e-d6e6-4726-ae09-673ec8916d05) Software Engineer – Platform Security job posting, San Francisco, 2026-03-27\n[P21](https://jobs.ashbyhq.com/friendliai/25429014-12bd-4665-a435-367d71a785c2) Software Engineer – Full Stack job posting, San Francisco, 2026-03-27\n[P22](https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362) Software Engineer – AI Agents job posting, San Francisco, 2026-04-15\n[P23](https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737) Software Engineer – Python Developer Tools job posting, San Francisco, 2026-04-15\n[P24](https://jobs.ashbyhq.com/friendliai/d5288f06-df73-407d-9d32-3eb96d735a51) Software Engineer – Senior Backend job posting, San Francisco, 2026-04-15\n[P25](https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7) Customer Success Engineer job posting, Seoul, 2026-04-17\n[P26](https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe) Software Engineer – AI Inference Engine job posting, Seoul, 2026-04-22\n[P27](https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735) Software Engineer – GPU Kernel job posting, Seoul, 2026-04-22\n[P28](https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09) Developer Advocate job posting, Seoul, 2026-04-24\n[W1](https://friendli.ai/blog/nvidia-nemotron-3-ultra) \"Nemotron 3 Ultra on FriendliAI\" blog post, 2026-06-04\n[W2](https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz) \"MiniMax-M3 on Friendli Dedicated Endpoints\" LinkedIn post, 2026-06-25\n[W3](https://friendli.ai/blog/kimi-k2-6) \"Kimi K2.6 Meets FriendliAI\" blog post, 2026-05-28\n[W4](https://friendli.ai/blog/deepseek-v4-pro-flash) \"DeepSeek V4 Pro and Flash on Dedicated Endpoints\" blog post, 2026-05-15\n[W5](https://friendli.ai/blog/GLM-5.2) \"GLM-5.2 Day-0 on FriendliAI Model APIs\" blog post, 2026-06-17\n[W6](https://friendli.ai/docs/guides/model-apis/introduction) \"Introducing Friendli Model APIs\" documentation, 2026-06-24\n[E4](https://github.com/friendliai/friendli-openapi) friendli-openapi repo creation event, 2026-06-22\n[E5](https://github.com/friendliai/friendli-openapi) friendli-openapi repo creation event, 2026-06-22\n[E27](https://github.com/friendliai/TensorRT-LLM) NVIDIA/TensorRT-LLM fork event, 2025-05-23\n[E28](https://github.com/friendliai/huggingface-blog) huggingface/blog fork event, 2025-01-22\n[E31](https://github.com/friendliai/lm-evaluation-harness) EleutherAI/lm-evaluation-harness fork event, 2023-08-18\n[E42](https://github.com/friendliai/pi-mono) earendil-works/pi fork event, 2026-02-23\n[E44](https://github.com/friendliai/opencode) anomalyco/opencode fork event, 2025-12-27\n[E46](https://github.com/friendliai/models.dev) anomalyco/models.dev fork event, 2025-12-23\n[E54](https://github.com/friendliai/flashinfer) flashinfer-ai/flashinfer fork event, 2025-07-09\n[E55](https://github.com/friendliai/cutlass) NVIDIA/cutlass fork event, 2025-07-08\n[E57](https://github.com/friendliai/tokasaurus) ScalingIntelligence/tokasaurus fork event, 2025-06-06\n[E58](https://github.com/friendliai/gorilla) ShishirPatil/gorilla fork event, 2025-01-14\n[E59](https://github.com/friendliai/simple-evals-archived) openai/simple-evals fork event, 2024-10-27\n[E60](https://github.com/friendliai/weaviate-io) weaviate/weaviate-io fork event, 2024-08-19","generated_at":"2026-06-27T19:06:03.505+00:00","citations":[{"url":"https://jobs.ashbyhq.com/friendliai/cb6e7c74-c22a-4192-aec8-65769839740b","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/cd4dd23b-cf94-46ec-afc3-f84037dca735","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.2","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-openapi","path":null,"label":"friendliai/friendli-openapi","type":"external"},{"url":"https://friendli.ai/blog/nvidia-nemotron-3-ultra","path":null,"label":"friendli.ai/blog","type":"external"},{"url":"https://www.linkedin.com/posts/friendliai_minimax-minimaxm3-aiinference-activity-7475940930437709824-JBwz","path":null,"label":"linkedin.com/posts","type":"external"},{"url":"https://friendli.ai/blog/kimi-k2-6","path":null,"label":"friendli.ai/blog","type":"external"},{"url":"https://friendli.ai/blog/deepseek-v4-pro-flash","path":null,"label":"friendli.ai/blog","type":"external"},{"url":"https://friendli.ai/blog/GLM-5.2","path":null,"label":"friendli.ai/blog","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/e5c4337b-338d-43e6-9bda-b82178acc3d1","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/f383e666-a99a-4574-b270-4e1738734706","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/b131cb30-8386-407e-a984-a75d5f2845da","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://github.com/friendliai/TensorRT-LLM","path":null,"label":"friendliai/TensorRT-LLM","type":"external"},{"url":"https://github.com/friendliai/flashinfer","path":null,"label":"friendliai/flashinfer","type":"external"},{"url":"https://github.com/friendliai/cutlass","path":null,"label":"friendliai/cutlass","type":"external"},{"url":"https://github.com/friendliai/lm-evaluation-harness","path":null,"label":"friendliai/lm-evaluation-harness","type":"external"},{"url":"https://github.com/friendliai/simple-evals-archived","path":null,"label":"friendliai/simple-evals-archived","type":"external"},{"url":"https://github.com/friendliai/langchain","path":null,"label":"friendliai/langchain","type":"external"},{"url":"https://github.com/friendliai/llama_index","path":null,"label":"friendliai/llama_index","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/eebbe221-65e4-4061-a519-8571219b6b7b","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/8da9914d-efe1-4bb1-96a9-a10eceab67fe","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/5fd918c3-70c8-44f1-b080-2aef9520b312","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/d5288f06-df73-407d-9d32-3eb96d735a51","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/25429014-12bd-4665-a435-367d71a785c2","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/41859c5f-fd60-40a3-9966-39a78852008b","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/2c922001-cccf-4006-9730-5488d4efaba6","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/6b7dbaf7-8751-402e-b253-ad968f7dc362","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/4cfe5d9f-d34f-4100-98b2-267c10238737","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/8d4ac792-7b6b-4503-84a0-a321ee68fe71","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/9df44597-ee03-4504-8bed-a842b6135b09","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/11f5ea91-9775-48a8-aca6-0438de737fd7","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/a6364382-e30a-4a30-919d-2c6b6c78185a","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://github.com/friendliai/gorilla","path":null,"label":"friendliai/gorilla","type":"external"},{"url":"https://github.com/friendliai/weaviate-io","path":null,"label":"friendliai/weaviate-io","type":"external"},{"url":"https://github.com/friendliai/locust_exporter","path":null,"label":"friendliai/locust_exporter","type":"external"},{"url":"https://github.com/friendliai/tokasaurus","path":null,"label":"friendliai/tokasaurus","type":"external"},{"url":"https://github.com/friendliai/pi-mono","path":null,"label":"friendliai/pi-mono","type":"external"},{"url":"https://github.com/friendliai/huggingface-blog","path":null,"label":"friendliai/huggingface-blog","type":"external"},{"url":"https://github.com/friendliai/opencode","path":null,"label":"friendliai/opencode","type":"external"},{"url":"https://github.com/friendliai/models.dev","path":null,"label":"friendliai/models.dev","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.0","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.1","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.6","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.8","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-openapi","path":null,"label":"friendliai/friendli-openapi","type":"external"},{"url":"https://github.com/friendliai/friendli-openapi","path":null,"label":"friendliai/friendli-openapi","type":"external"},{"url":"https://friendli.ai/docs/guides/model-apis/introduction","path":null,"label":"friendli.ai/docs","type":"external"},{"url":"https://jobs.ashbyhq.com/friendliai/372dbe1e-d6e6-4726-ae09-673ec8916d05","path":null,"label":"jobs.ashbyhq.com/friendliai","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.10.5","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.10.6","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.10.7","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.10.9","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.11.0","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.1","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.2","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.3","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.4","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.6","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.12.8","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.0","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.1","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-python/releases/tag/v0.13.2","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/friendli-client","path":null,"label":"friendliai/friendli-client","type":"external"},{"url":"https://github.com/friendliai/periflow-cli","path":null,"label":"friendliai/periflow-cli","type":"external"},{"url":"https://github.com/friendliai/FAI-Model","path":null,"label":"friendliai/FAI-Model","type":"external"},{"url":"https://github.com/friendliai/friendli-model-optimizer","path":null,"label":"friendliai/friendli-model-optimizer","type":"external"},{"url":"https://github.com/friendliai/LLMServingPerfEvaluator","path":null,"label":"friendliai/LLMServingPerfEvaluator","type":"external"},{"url":"https://github.com/friendliai/aipm","path":null,"label":"friendliai/aipm","type":"external"},{"url":"https://github.com/friendliai/friendli-gradio","path":null,"label":"friendliai/friendli-gradio","type":"external"},{"url":"https://github.com/friendliai/examples","path":null,"label":"friendliai/examples","type":"external"},{"url":"https://github.com/friendliai/llm-hackathon-tutorial","path":null,"label":"friendliai/llm-hackathon-tutorial","type":"external"},{"url":"https://github.com/friendliai/friendli-python","path":null,"label":"friendliai/friendli-python","type":"external"},{"url":"https://github.com/friendliai/lm-evaluation-harness","path":null,"label":"friendliai/lm-evaluation-harness","type":"external"}],"provenance":{"provider":"deepseek","model":"deepseek-v4-pro","workflow":"onlylabs-deepagents-analysis-v3","agent":"deepagents"},"evidence":{"total":94,"pages":28,"events":131,"web":6,"signal_desks":{"forks":12,"repos":14,"hiring":20,"talking":0,"releases":14},"data_radar_lanes":null,"data_radar_matches":null}},"signal_counts":{"total":131,"model_released":0,"release":75,"repo_new":17,"repo_forked":19,"post_published":0,"job_opened":20}}