Location: [India]
Reports to: MDM & Data Governance CoE Lead / Practice Lead
Team: AI/Agentic Engineering, Commercial Data Solutions
We're looking for an AI Developer to build and productionize agentic AI, LLM-based, and applied ML solutions that power commercial data platforms (MDM, data governance, HCP/HCO mastering, customer 360) for life sciences clients. This is a hands-on engineering role — you'll design, code, evaluate, and deploy AI systems, not just prototype them. Pharma domain knowledge is a plus, not a prerequisite; we'll teach you the industry context. What matters is that you can ship reliable AI products.
- Design and build LLM-powered agents (RAG pipelines, tool-using agents, multi-agent orchestration) for use cases like entity resolution, data stewardship automation, hierarchy/relationship inference, and natural-language querying over master data
- Develop and fine-tune models (prompt engineering, fine-tuning, embeddings, evaluation harnesses) and integrate them into production pipelines
- Build robust evaluation frameworks — accuracy, hallucination rate, latency, cost — and iterate against them
- Own the full lifecycle: architecture, coding, testing, deployment, monitoring, and versioning of AI components
- Integrate AI services with enterprise data platforms (Informatica IDMC, Reltio, Databricks, Snowflake, etc.) via APIs and MCP-style tool connectors
- Collaborate with solution architects and delivery leads to translate client requirements (e.g., HCP affiliation logic, consent management, account hierarchy) into AI-driven technical designs
- Build internal accelerators/reusable agent frameworks that can be deployed across multiple client engagements
- Contribute to technical proposals and PoCs during pre-sales when needed
- Strong Python engineering skills (not just notebooks — production code, testing, packaging)
- Hands-on experience with LLM APIs (OpenAI, Anthropic, etc.) and orchestration frameworks (LangChain, LangGraph, LlamaIndex, or equivalent)
- Practical experience building RAG systems: chunking strategies, vector databases (Pinecone, Weaviate, pgvector, etc.), retrieval evaluation
- Understanding of agentic architectures: tool calling, function calling, multi-step planning, memory/state management
- Experience with prompt engineering and systematic prompt evaluation (not trial-and-error)
- Familiarity with model fine-tuning and/or embedding model selection/tuning
- API development (REST, sometimes MCP) and integration with third-party/enterprise systems
- Comfortable with cloud platforms (Azure/AWS/GCP) and basic MLOps (CI/CD for models, monitoring, logging)
- SQL and experience working with structured/semi-structured enterprise data
- Exposure to master data management concepts (match/merge, golden records, survivorship) — will be trained if not present
- Experience with Databricks, Snowflake, or similar data platforms
- Familiarity with life sciences/pharma commercial data (HCP/HCO, territory, consent) — helpful but not required
- Experience building internal tools/accelerators reused across projects
- Contributions to open-source AI tooling
