Data Scientist

Years of experience: 6- 12 years

Locations: Bangalore, Noida, Indore

• 6-10 years of software engineering experience, with at least 2+ years building production AI/ML or LLM-powered applications.
• Deep expertise in LangChain and LangGraph (you have built non-trivial multi-agent systems with cycles, persistence, and streaming).
• Strong experience with MCP (Model Context Protocol) and A2A (Agent-to-Agent) communication standards.
• Expert-level Python, FastAPI, async programming, WebSockets, and building scalable APIs.
• Proven track record designing and shipping microservices or large-scale enterprise applications.
• Hands-on experience deploying containerized workloads on GKE (or equivalent like EKS/AKS) – Helm, Kustomize, GitOps (ArgoCD/Flux), IAM, networking.
• Solid understanding of LLM evaluation, prompt engineering, retrieval-augmented generation (RAG), and agent reliability techniques.
• Experience with observability in AI systems (LangSmith, Phoenix, Helicone, or custom tracing).
• Strong grasp of software engineering fundamentals: testing (pytest, integration tests for agents), CI/CD pipelines, design patterns, Code reviews etc.
• Experience working on Agile delivery methodology.

Roles & Responsibilities

• Design and build reusable Agentic AI framework and workflows on top of LangChain/LangGraph (VertexAI or Bedrock)
• Lead integration of agents with external tools/APIs via Model Context Protocol (MCP), Agent-to-Agent (A2A) protocol, function calling, and custom connectors.
• Build high-performance backend services using FastAPI, async Python, Pydantic, and event-driven patterns.
• Design systems following microservices and enterprise application design principles.
• Ensure automated CI/CD pipelines and production deployment on Google Kubernetes Engine (GKE), including autoscaling, observability (Prometheus/Grafana, OpenTelemetry), and resilience patterns.
• Awareness of using LangSmith, LangFlow or similar such frameworks
• Establish engineering best practices for agent development (testing agents, version control of prompts/tools, rollback strategies, guardrails).
• Collaborate with product, data science, and platform teams to productionize agents that solve real enterprise problems.
• Expertise in writing clear and concise prompts with different tuning mechanism
• Expertise in implementing RAG based solution/agents leveraging knowledge base creation using vector databases.
• Mentor junior engineers and conduct architecture & code reviews.
• Well equipped in using Code assist tools like copilot, cursor, windsurf or other equivalents.