5+ years of software engineering experience, with 3+ years in production LLMs or ML.
Proven track record of delivering agentic systems relied upon by users.
Extensive experience in prompt/context engineering, tool/API design, and vector DBs.
Strong grasp of model behavior, including fine-tuning and latency/cost tradeoffs.
Experience in setting technical direction across multiple teams.
Ability to communicate complex systems clearly and simply.
Bonus: Experience with high concurrency microservices for agent workloads.
Bonus: Familiarity with GLiNER or similar architectures.
Responsibilities
Design and own the architecture for agentic systems on Fastino's models.
Identify where specialized models outperform frontier LLMs in agent loops and validate with benchmarks.
Develop evaluations, telemetry, and feedback loops to assess agent performance.
Collaborate with research to translate new model architectures into agent capabilities.
Ensure agents are fast and reliable at scale, focusing on low latency and high concurrency.
Drive data and fine-tuning strategies for agent-specific tasks like tool calling and routing.
Benefits
Flexible work environment in a vibrant San Francisco location.
Opportunity to work on cutting-edge technology in AI and ML.
Collaborative culture that encourages innovation and knowledge sharing.
Access to professional development resources and opportunities.
Potential for involvement in open-source projects and research contributions.
Full Job Description
Distinguished Agent Engineer
Full-time | San Francisco
The Role
This is one of our most senior IC engineering role. You'll set the technical direction for how agents get built on Fastino models, write a lot of the core code yourself, and raise the bar for everyone around you. You've built agentic systems that real users depend on, and you have strong opinions about what works.
What You'll Work On
Design and own the architecture for agentic systems built on Fastino's small, fast models: planning, tool use, memory, and multi-agent orchestration.
Figure out where specialized models beat frontier LLMs inside agent loops, and prove it with benchmarks and real customer workloads.
Build the evals, telemetry, and feedback loops that tell us when agents are getting better (or worse) and why.
Partner with research to turn new model architectures into agent capabilities, and shape what research works on next.
Make agents fast and reliable at scale: low latency inference, high concurrency serving, graceful failure handling.
Drive data and fine-tuning strategy for agent specific tasks like tool calling, extraction, and routing.
What We're Looking For
5+ years of software engineering experience, with 3+ years building with LLMs or ML in production.
A track record of shipping agentic systems that real users rely on, and the scars to prove it.
Deep experience across the stack: prompt and context engineering, tool and API design, vector DBs, Kubernetes, and inference serving.
Strong understanding of model behavior: fine-tuning, evals, failure modes, and latency/cost tradeoffs.
Experience setting technical direction across teams, not just within one project.
You write clearly and make complicated systems feel simple.
Bonus: experience building high concurrency microservices for agent workloads.
Bonus: familiarity with GLiNER or other information extraction architectures.
Bonus: open source contributions, papers, or talks on agents or LLM systems.
