Role & responsibilities

Senior Generative & Agentic AI Engineer

Experience: 58 years (at least 2 years hands-on with LLMs and agents)

Location: Bengaluru / Hybrid Track: Generative AI & Agentic Systems

Role Summary

We are hiring a Senior GenAI Engineer to design and build agentic applications — multi-step, tool-using LLM systems that solve real enterprise workflows. You will own the agent layer end-to-end: prompt design, orchestration graphs, tool integration, evaluation, and production hardening. This role is for engineers who have moved beyond chat wrappers and have shipped LangChain/LangGraph agents on a managed cloud LLM platform in production.

What You'll Do

- Design and build agentic applications with LangChain and LangGraph — multi-agent workflows with planning, tool use, memory, and human-in-the-loop checkpoints.
- Build and operate solutions on AWS Bedrock, Azure AI Foundry, or Google Vertex AI — model selection, invocation, streaming, guardrails, and platform-native agent capabilities (Bedrock Agents/AgentCore, Foundry Agent Service, Vertex AI Agent Builder).
- Engineer RAG pipelines — chunking strategies, embeddings, hybrid retrieval, reranking, and grounding — using the chosen cloud's managed services or OSS equivalents.
- Build and integrate tools/functions that agents call (APIs, databases, MCP servers, internal systems).
- Define and run evaluation harnesses — golden sets, LLM-as-judge, RAGAS-style metrics, regression suites.
- Implement guardrails: input/output validation, PII handling, prompt-injection defenses, cost and latency controls.
- Collaborate with the Python platform team to productionize agents reliably; partner with product and domain experts on use-case framing.




- Stay current with the model and tooling landscape, and advise on build-vs-buy and model-routing decisions.

Must-Have Skills

- 5–8 years of overall engineering experience, with solid Python fluency.
- 2+ years hands-on with LLM application development — beyond chatbot demos.
- Mandatory: production experience with LangChain and LangGraph — built, debugged, and shipped agents using both. LangGraph state machines, checkpointers, and tool-node patterns should be familiar territory.
- Mandatory: hands-on production experience on at least one of — AWS Bedrock, Azure AI Foundry (incl. Azure OpenAI), or Google Vertex AI. Must have used the platform's native model APIs, agent/orchestration features, and guardrail capabilities — not just called the underlying models.
- Working knowledge of RAG architectures and at least one vector database (Pinecone, Weaviate, pgvector, OpenSearch, FAISS, Milvus, or the chosen cloud's managed equivalent).
- Practical command of prompt engineering, function/tool calling, structured outputs (JSON schema), and context-window management.
- Familiarity with evaluation and observability for LLM apps (RAGAS, LangSmith, Langfuse, Phoenix, or custom harnesses).
- Awareness of emerging protocols — MCP, A2A, or similar — and how agents integrate with enterprise tools.

Nice-to-Have

- Experience with streaming, async orchestration, and long-running agent execution patterns.
- Exposure to a second cloud's AI platform (useful, not required).
- Exposure to fine-tuning, LoRA/QLoRA, or model distillation.
- Background in regulated domains (BFSI, healthcare) and familiarity with AI governance frameworks.
- Open-source contributions or technical writing in the GenAI space.

Preferred candidate profile

📌 Walk-in || Python+ Generative AI (Bengaluru)
🏢 LanceSoft
📍 Bengaluru