At a glance
- Experience
- 1-4 years experience
- Employment
- Full-time
- Workplace
- Hybrid
- Location
- Pune
About the role
- An AI Engineer independently designs and delivers production AI systems, owning the behavior and quality of AI applications.
- The role encompasses agent architecture, orchestration, memory, model selection, evaluation, reliability, and operational tradeoffs.
- This position also contributes to building and scaling AI and agent platforms that serve as a harness for non-AI engineers.
Responsibilities
- Build, deliver, and maintain production-grade LLM, agent, retrieval, and machine-learning workflows, product platforms, and features using structured orchestration patterns.
- Design, define, and maintain agent graphs, MCPs, states, transitions, routing logic, retries, fallbacks, escalation behavior, and tool integrations using established patterns.
- Develop and refine comprehensive prompting strategies across system prompts, task prompts, tool instructions, response schemas, guardrails, and multi-agent interactions.
- Configure and design short-term state, long-term memory, conversation history, summarization, retrieval, and privacy boundaries.
- Select and configure appropriate models based on task requirements, capability, latency, cost, context length, operational constraints, and reliability.
- Use AWS AI services and platform capabilities, such as Amazon Bedrock and AgentCore, to build, deploy, and operate AI systems.
- Design robust retrieval and grounding strategies, including chunking, embeddings, ranking, metadata, filtering, and source attribution.
- Create, run, and scale quantitative/qualitative evaluation suites that measure correctness, completeness, reasoning, safety, tool use, latency, grounding, and regression detection.
- Investigate and analyze unexpected agent behavior, evaluation results, and production traces to diagnose failures.
- Own AI projects from technical design through production deployment and ongoing improvement while providing technical guidance and code reviews.
Requirements
- 1–4 years of experience in software engineering or related fields, including 2+ years with AI, data science, or related technologies.
- Strong proficiency in Python, JavaScript, TypeScript, or a similar programming language.
- Strong working knowledge of Node.js and JavaScript or TypeScript.
- Demonstrated experience building and operating LLM or agentic systems in production.
- Strong understanding of prompting, tool calling, structured outputs, agent orchestration, state management, context construction, and memory.
- Experience designing agent graphs, multi-step workflows, routing strategies, and failure-handling behavior.
- Experience selecting and evaluating models for different tasks and understanding quality, latency, cost, and reliability tradeoffs.
- Practical experience with Amazon Bedrock, AgentCore, or equivalent cloud AI services.
- Experience with RAG systems, vector databases, embeddings, ranking, and grounding techniques.
- Strong system design, communication, ownership, and collaboration skills.
Skills
- Python
- JavaScript
- TypeScript
- Node.js
- AWS
- Amazon Bedrock
- AgentCore
- LLM
- RAG
- Vector Databases
