Software Engineer (Middle+) — Python/FastAPI + React

About the role
We are looking for a Software Engineer to join our team working on AI-powered pharma industry products. You will work across two core platform products — building backend services, APIs, and AI-powered features, and shipping the frontend side of your work in React + TypeScript. You will work closely with the product owner, frontend developers, QA, DevOps, and biologists.

Must-haves

• A completed university degree in Computer Science, Software Engineering, or a related technical field (Bachelor’s or higher).
• Strong algorithmic thinking — solid data-structures foundation, able to reason about complexity and design correct, efficient solutions rather than only wiring frameworks.
• ~4+ years of production experience building backend services in Python (FastAPI or a similar async framework) with PostgreSQL / SQLAlchemy.
• Hands-on experience with AI-assisted / agentic development (Claude Code or similar) — and the ability to configure and tune these tools, not just use them out of the box; you can explain where you trust them and how you verify their output.
• Practical experience with MCP (Model Context Protocol) and RAG (retrieval-augmented generation, embeddings, vector search).
• Comfortable shipping features in React + TypeScript — you can build and wire up UI without a frontend engineer holding your hand.
• Solid with Docker and Git-based workflows; understand how services run in Kubernetes.

• English(B1-B2) sufficient to work in an English-language codebase and team.

   

Tech Stack

• Core: Python 3.10+, FastAPI, async/await, SQLAlchemy 2.0, Alembic
• Frontend: React 18, TypeScript, Vite
• Database: PostgreSQL, DuckDB, vector search (embeddings)
• AI/LLM: Claude Code, MCP (Model Context Protocol), RAG pipelines
• Infrastructure: Docker, Kubernetes, Redis (FastStream)
• Storage: Parquet
• Orchestration: Temporal, Argo (nice to have)
• CI/CD: GitOps, ArgoCD, GitLab CI
• Observability: structured logging, metrics, distributed tracing

• Tools: Git, Jira, Confluence

   

Nice-to-have (not required)

• Experience with workflow orchestration (Temporal, Argo, Airflow) or message-driven architectures.
• Building your own MCP servers, or productionising LLM apps (evals, guardrails, cost/latency tuning).
• Data/analytics tooling (DuckDB, Parquet) or a scientific/chemistry background.

• GitOps / CI experience (ArgoCD, GitLab CI).

   

What you’ll do

• Design and ship async REST APIs in FastAPI so researchers can run screening, manage projects, and explore results.
• Integrate backend services with our compute-orchestration layer so long-running docking and screening jobs are durable and observable.
• Drive features with an AI-first workflow: plan and implement with Claude Code, then critically review, test, and verify what the agent produces before it ships.
• Model and evolve the PostgreSQL schema (SQLAlchemy 2.0, Alembic), shipping safe migrations as the product grows — with the team’s support on the trickier zero-downtime changes.
• Build and integrate the AI layer — MCP tools and RAG pipelines (embeddings, vector search) that power our in-product AI Agent — with evals, guardrails, and retrieval quality treated as part of the feature, because the agent informs scientific decisions.
• Build async, event-driven services: Redis/FastStream for lightweight service-to-service events
• Ship the frontend side of your features in React 18 + TypeScript (Vite) — wiring UI to the APIs you build
• Keep your services observable and reliable in production: structured logging, metrics, and distributed tracing across async services and Temporal workflows
• Review peers’ code — including AI-generated code — as part of how the team keeps quality high.
   

Domain knowledge 

• Basic understanding of molecular biology — proteins, DNA/RNA, small molecules
• Experience in biotech, pharma, or life sciences software is a strong plus
  
  Soft skills
• Attention to detail
• Proactive communication — if something is unclear, asks immediately
• Ability to work independently and manage own workload