Job Description

ABOUT THE ROLE

We are building a new product at the intersection of agentic AI and hard backend engineering. The work runs deep: distributed systems, demanding data problems, and AI engineering that takes large language models beyond demos, into systems you can trust.

As a Senior AI Developer you will be the team’s reference on LLMs and agentic systems: you design the agents, the retrieval and evaluation pipelines, and the tooling that turns models into dependable product features — and you help define how the whole team practices AI.

WHAT YOU’LL DO

• Design and deliver AI features end to end: agent orchestration, tool calling, RAG pipelines, implementation, tests, and production rollout.
• Own the hardest AI problems: reliability, evaluation, cost and latency, and the guardrails that make LLM-powered features trustworthy.
• Build the evaluation harnesses and datasets that tell us whether an agent is actually getting better.
• Lead delivery with junior developers: break features into tasks, review their code, and raise the bar.
• Track the model and framework landscape, turn what matters into proofs of concept, and proofs of concept into product.
• Document the designs and decisions that keep the codebase evolvable.

WHAT WE EXPECT FROM YOU

• Own outcomes: when you take a feature, it ships — designed well, tested, and supported in production.
• Be the LLM expert: the team’s reference for models, agent patterns, and their limits — and for when not to use them.
• Drive technical excellence: set, refine, and enforce the standards the rest of the team builds on.
• Multiply the team: mentor junior developers and give feedback that makes people better, not just code.
• Think in abstractions: our codebase is built on a few deep concepts — you extend them without breaking them.

WHAT YOU BRING

• Deep hands-on LLM / agentic AI experience: agent orchestration, tool calling, RAG, prompt design, and structured evaluation — in production, not just notebooks.
• Strong Python plus solid Java (or another strong object-oriented language) — our agents run inside a hard-engineering backend.
• Experience integrating models through provider APIs and frameworks (LangChain4j, LangChain, or similar), including streaming, tool/function calling, and context management.
• A grip on the failure modes — hallucination, prompt injection, drift — and the evaluation and guardrail techniques that contain them.
• Production discipline: thorough testing, CI, observability, and cost/latency awareness for model-backed features.
• A track record of mentoring developers and leading feature delivery.

BONUS POINTS

• Fine-tuning, distillation, or serving open-weight models (LoRA, vLLM, Ollama).
• Agent ecosystems: MCP, multi-agent orchestration, structured agent evaluation.
• Vector search and embedding pipelines (pgvector, Qdrant, Elasticsearch, or similar).
• Distributed-systems background: REST/RPC services, caching, multi-tenancy.
• Enough React/TypeScript to be dangerous.

QUALIFICATIONS

• Bachelor degree in Computer Science or Computer Engineering.
• Other degrees considered with a proven track record in software development.