We build secure, source-grounded AI agents for industrial companies. Your job: design, chain, and ship them end-to-end.

What you’ll do

• Own solution design: pick the right model and tools for the job (reasoning, extraction, tool-use, reranking), compare vendors, run bake-offs, and explain trade-offs.

• Engineer prompts & chains: system prompts, tool/function calling, JSON-schema outputs, guardrails, and multi-step orchestration.

• Build the glue software: Python/TypeScript services (FastAPI/Node), Docker, CI/CD, observability (latency, cost, traces).

• Ship production RAG: chunking/segmentation, hybrid search (BM25+embeddings), rerankers, vector/graph stores, caching, and eval harnesses.

• Orchestrate workflows: use our agent builder to build the agent skeleton, configure webhooks, and connectors into CRM/ERP/PLM and data lakes.

• Enforce safety & security: authZ/authN, least-privilege data access, PII minimization, audit logs, red-team tests.

• Partner with customers: turn messy requirements into working agents; measure impact and iterate quickly.

What you bring

• Fluency selecting and using frontier and open-source LLMs (OpenAI, Anthropic, Google, Meta, Mistral, etc.) plus embeddings, re-rankers, and small task-specific models.

• Strong prompt design skills and practical chaining/orchestration experience (Langflow, LangChain/LlamaIndex or equivalent—plus your own lightweight patterns).

• Hands-on RAG builds in production (not just demos): data pipelines, evals (e.g., RAGAS/DeepEval/custom), and quality/latency/cost trade-offs.

• Comfortable writing real software: Python/TypeScript, APIs, testing, Docker, cloud (AWS/GCP/Azure).

• Workflow tooling: you’ve automated with n8n, Zapier or other agent and workflow building tools and know when to replace them with code.

• Up-to-date on the AI tool landscape; you track releases, run quick experiments, and keep us current.

Nice to have

• UI chops (Next.js/React) for simple operator/customer surfaces.

• Data/connectors: SQL, REST, GraphQL, message queues.

• Industry context in manufacturing/industrial automation (PLCs, SCADA, servos, motors, etc) a plus.

How we work

• Small team, high ownership. You build, ship, and measure.

• We optimize for reliable, verifiable answers, grounded in the customer’s data.

• Fast iterations, real users, clear KPIs.