You'll build the agent layer that powers Dynabase — orchestration, memory, tool-use, and evaluation. This is close-to-model work on a product that runs real business operations.
- Design and ship the orchestration layer for Dynabase agents
- Build memory and context systems that make agents reliable over time
- Implement tool-use pipelines for accounting, compliance, and ERP integrations
- Own evaluation frameworks to measure agent accuracy and reliability
- 3+ years of backend engineering in Python or TypeScript
- Hands-on experience building with LLMs — not demos, production systems
- Robust fundamentals in distributed systems and data pipelines
- Comfortable working close to the model and close to the business logic
- Experience with LangChain, LlamaIndex, or similar frameworks
- Knowledge of Indian financial or compliance data systems
Apply for this role
Send us your details and a CV. Optionally, tell us what you've shipped that you're proud of — we read every one.