Solution design & backbone engineering- Own end to end GenAI pipelines for CSR automation: Ingestion à preprocessing/OCR à semantic chunking à embeddings/vector store à RAG/agentsà evaluation & observability.
Engineer agentic workflows with LangGraph / AutoGen / CrewAI / Agno to orchestrate multi step reasoning, tool use, and human checkpoints.
Modularize codebases to support pluggable models, prompts, retrievers, and knowledge graphs across content creation, summarization, and Q&A.;
Data, retrieval, and knowledge graphs- Connect to Databricks for data engineering and feature pipelines
Implement semantic chunking/tokenization strategies for clinical artifacts (TLFs, SAP, protocol, listings) and optimize retrievers for numerical fidelity.
Stand up and tune vector databases; design Knowledge Graphs (e.g., Neo4j) and Graph RAG patterns for evidence and traceability.
Establish ground truth/eval sets and automatic scoring for accuracy, completeness, and faithfulness.
Modeling, prompts, and quality- Select and integrate LLMs via Azure OpenAI/AWS Bedrock




Design prompt templates and function/tool schemas for consistent outputs.
Build guardrails and evaluators to reduce hallucination
Deployment & MLOps (Valuable to have)- Containerize and deploy services with Docker on Azure/AWS; implement CI/CD, secrets, model/version registries, and cost observability.
Integrate low/no code orchestrators (n8n) for human workflows and rapid business iteration
Maintain production reliability (latency, cost per page/section, eval scores) and security (PII handling, audit logs)

Required Skills and Qualification:
Advanced degree in computer sciences or quantitative research or relevant experience in AI/ML space with hands-on experience in end-to-end (development to deployment)
5+ yrs AI/ML engineering (2+ yrs Generative AI in production).
Robust Python; experience building back-end services and APIs (bonus).
Familiarity with OCR technologies and multi-modal data proc

📌 Project Leader Hinjawadi
🏢 Axtria
📍 Hinjawadi