About the role
You'll architect and ship across the full Genesis stack: agentic pipelines, backend APIs, data infrastructure, and clinical-facing UI. You'll work directly with founders and customers. You'll own things end-to-end. This is not a role where you bolt AI onto existing CRUD. You'll be making foundational decisions about how intelligent systems are designed, evaluated, and operated at scale in a regulated industry.
The core requirements for the job include the following:
You build robust backend systems:
- 6 + years building production web applications from scratch.
- Deep Python proficiency; comfortable with FastAPI, Django, or Flask in production.
- Experience designing APIs that serve both humans and AI agents (tool schemas, structured outputs, streaming).
- Async-first thinking: asyncio, task queues, event-driven architectures.
- Good to have Kafka, Redis, or ActiveMQ for real-time data movement.
- Nice to have Postgres, Elasticsearch, MongoDB, or graph databases (Neo4j, TigerGraph) in production.
AI-native engineering is your default mode:
- You've built production systems where LLMs are doing real work - not demos, not PoCs.
- You've designed and shipped RAG pipelines, multi-agent workflows, or tool-using agents in production.
- You understand prompt engineering as an engineering discipline: versioning, evaluation, and regression testing.
- You've instrumented AI systems for observability - latency, token usage, hallucination rate, and drift.
- You can reason about model tradeoffs (context length, cost, latency, accuracy) and make architectural calls accordingly.
- You've worked with LLM SDKs (OpenAI, Anthropic, Bedrock, etc. ) and agentic orchestration frameworks (LangChain, LlamaIndex, CrewAI, or similar).
You operate at cloud scale:
- Docker and Kubernetes in production - this is a hard requirement.
- At least one public cloud (AWS, Azure, GCP) with real operational experience.
- Microservices and cloud-native design patterns.
- You've been on-call. You know what a bad deploy feels like at 2 am.
- You can ship a frontend when the product demands it.
- Nice to have React, TypeScript, or modern JS frameworks.
- Enough frontend fluency to build clinical interfaces without a dedicated frontend handoff.
Bonus Points:
- Led a small engineering team - mentored, reviewed, unblocked.
- CKAD or equivalent Kubernetes certification.
- ML/DL model deployment experience (PyTorch, scikit-learn).
- Built evaluation harnesses or used MLflow, LangSmith, or similar for AI observability.
- Healthcare domain experience (FHIR, HL7 clinical workflows).