About the job AI Engineer
About the role
The AI Engineer builds and ships Mimoid software using agentic development workflows. You take product requirements, drive coding agents and frameworks to implement against them, verify the resulting software actually meets the requirements, manage releases, and deploy into the cloud.
What you'll do
- Build with agents: develop software using agentic coding workflows and frameworks — orchestrating coding agents to implement features across Mimoid's stack rather than hand-writing every line.
- Build the domain agents: contribute to Mimoid's agent roster (EBOMMBOM, work packages, scheduling, change-impact, flow advisor, quality, progress-vision), each running graph context model write-back with human-in-the-loop gates and an eval-gated prompt registry.
- Own requirements-to-code: understand requirements from the Solutions Lead and product; decompose them; and develop/guide agents to implement them correctly.
- Verify quality: write and run evals and tests, review agent output, and catch regressions and hallucinations before release — trust tiers, calibration, and no pass, no go.
- Manage releases: own versioning, release management, and change control across the FastAPI API and Celery worker services.
- Deploy to cloud: containerize and ship into the cloud — CI/CD, migrations, environment config, and post-deploy verification, observability, and cost-per-build telemetry.
Core tech stack (familiarity required)
Agentic dev
Coding-agent workflows & frameworks (e.g., Claude Code / Cursor / agent SDKs; LangGraph, LangChain, CrewAI, DSPy); prompt engineering; evals
Backend
Python 3.11, FastAPI, Celery, Pydantic v2, SQLAlchemy 2.x, Alembic
Front end
React 18 + TypeScript, TanStack React Query, Tailwind / shadcn UI, Vite
AI / models
Managed frontier LLMs via a cloud model provider, tiered by capability (small medium large); RAG, local embeddings, model routing by complexity, eval-gated prompt registry
Data
PostgreSQL + pgvector, Neo4j (knowledge graph), object store, Redis; pandas; OR-Tools & networkx (scheduling engine)
Cloud / delivery
Docker containers, cloud infrastructure, CI/CD, migrations, observability, cost-per-build instrumentation
Required qualifications
- 4+ years building and shipping production software, with hands-on experience using AI/agentic tooling to develop and deliver code.
- Full-stack competence across a modern Python + TypeScript stack (FastAPI and/or Node/Express, React).
- Practical experience with LLM application patterns — RAG, agents, tool-calling, and prompt/eval workflows.
- Solid grasp of requirements analysis, testing/evals, release management, and cloud deployment (containers, CI/CD, migrations, environments).
- Comfort in a fast, customer-driven forward-deployed environment, occasionally on-site.
Preferred qualifications
- Experience with knowledge graphs (Neo4j) and vector retrieval (pgvector).
- Model-serving on a managed cloud provider; local/open-weight models and LoRA fine-tuning.
- Optimization/scheduling (OR-Tools CP-SAT, critical-path methods), PLM/BOM, or planning software.
- Complex, high-mix / low-volume manufacturing or PLM/BOM domain exposure.