Build and ship production agentic AI features — agents, tools, prompts, evals, and integrations — against an established reference architecture.
Required Qualifications:
• 3–8 years of experience in software development or data engineering
• Hands-on experience in Generative AI or LLM-based applications
• Experience building APIs, microservices, or distributed systems
• Bachelor's or Master's degree in Computer Science, AI/ML, Data Science, or related field
Key roles:
• Implement agents and sub-agents (planner, executor, critic, router) using Claude Agent SDK / Lang Graph
• Build tools and MCP integrations, design clean tool schemas, idempotent operations, and robust error handling.
• Implement RAG pipelines: ingestion, chunking, embedding (Bedrock Titan), hybrid retrieval, citation rendering.
• Develop Fast API/Python services exposing agent capabilities (sync + streaming); integrate with SQL (Postgres) and object stores (S3).
• Write evaluation harnesses (golden sets, regression suites, LLM-as-judge) and trace/observe agent runs.
• Implement guardrails: input/output validation, schema enforcement, rate limiting, prompt-injection defenses.
• Participate in code reviews, pairing, and architecture discussions; own quality of the code you ship.
• Strong Python (FastAPI, async, Pydantic) or Node/TypeScript equivalent.
• Hands-on with at least one agent framework (Claude Agent SDK / Lang Graph / AutoGen).
• Practical experience with LLM tool/function calling, structured outputs, streaming.
• RAG implementation experience (pgvector / FAISS / OpenSearch).
•Git, CI/CD, containerization (Docker), and cloud basics (AWS preferred).
Roles/Responsibilities:
• Implement single-agent and multi-agent systems using frameworks such as LangChain, Semantic Kernel, CrewAI, AutoGen, or similar
• Build applications using LLMs (Azure OpenAI, OpenAI, Anthropic, etc.)
• Implement Retrieval-Augmented Generation (RAG) pipelines
• Enable agents to coordinate and collaborate in multi-agent ecosystems
• Build secure, scalable APIs and microservices to support AI agents
• Develop evaluation frameworks for agent performance (accuracy, hallucination detection, response quality)
• Monitor system behavior and continuously improve reliability
• Optimize performance for latency, cost, and scalability