Software Enginee
Posted Jun 28 · 0 applicants
Saving, applying or scoring takes a few seconds to set up your free account.
The role in plain words
- Strong software engineering fundamentals: testing, type safety, system design.
- Strong Python.
- Shipped applications backed by language models with real multi-step pipelines, not just API calls.
- RAG end-to-end: embedding, retrieval, reranking, generation; vector and/or graph databases (Neo4j, Pinecone, Qdrant).
- Comfortable across the stack: you can build and ship a React/TypeScript frontend, not just an API.
- gRPC / protobuf services.
- Experience with edge and on-prem systems; air-gapped delivery.
- Industrial / mission-critical domain exposure (sensor data, P&IDs, protocols like Modbus/OPC UA).
- Foundation-model adaptation (SFT, LoRA), document/OCR or multimodal pipelines.
- Real-time data viz or operator/monitoring dashboards.
Extracted from the job description · kept up to date automatically
Who this suits
Full job description
Original listing · kept for referenceOverview We build AI systems that predict and reason about physical industrial systems, and ship them on-premise where being wrong has real consequences: power, critical infrastructure, defense.
This is a high-agency role on a small engineering team, working in close synergy with our research group. You own complete subsystems, not tickets, from the backend that runs our models to the UI operators rely on. We see this role as a strong software engineer who knows how to build systems with non-deterministic components, and engineer everything around them so the whole thing behaves predictably in production. You won't be doing research yourself, and you won't be managing infra. You partner with the researchers building our models and turn their work into the product that gets shipped to customers and operated under conditions where mistakes are expensive.
What you'll own • Production AI systems: agentic, multi-step pipelines combining language models (LLM/SLM), retrieval, and tool use into coherent runtimes.
• Retrieval: graph and vector retrieval, reranking, query planning (Neo4j knowledge graph, timeseries).
• The product around the model: FastAPI REST + WebSocket backend and the Next.js / React / TypeScript frontend operators actually use.
• Production hardening: testing, type safety, observability, evaluation, including regression suites and benchmarks so model behavior doesn't drift silently.
Requirements • Strong software engineering fundamentals: testing, type safety, system design. Strong Python.
• Shipped applications backed by language models with real multi-step pipelines, not just API calls.
• RAG end-to-end: embedding, retrieval, reranking, generation; vector and/or graph databases (Neo4j, Pinecone, Qdrant).
• Comfortable across the stack: you can build and ship a React/TypeScript frontend, not just an API.
• Pragmatic about tech debt: you fix what matters, not what bothers you aesthetically.
Nice to have • gRPC / protobuf services.
• Experience with edge and on-prem systems; air-gapped delivery.
• Industrial / mission-critical domain exposure (sensor data, P&IDs, protocols like Modbus/OPC UA).
• Foundation-model adaptation (SFT, LoRA), document/OCR or multimodal pipelines.
• Real-time data viz or operator/monitoring dashboards.
Questions about this role
- This listing did not state a salary. We only show pay when the employer publishes it.
- Strong software engineering fundamentals: testing, type safety, system design., Strong Python., Shipped applications backed by language models with real multi-step pipelines, not just API calls., RAG end-to-end: embedding, retrieval, reranking, generation; vector and/or graph databases (Neo4j, Pinecone, Qdrant)., Comfortable across the stack: you can build and ship a React/TypeScript frontend, not just an API.