Agentic AI Engineer Job
yash
Job Description
- Agent Architecture & Development
- Design and implement agentic AI architectures, including:
- ReAct (Reason + Act)
- Plan‑and‑Execute
- Tool‑using agents
- Multi‑agent systems and coordinators
- Build LLM‑powered applications using modern foundation models
- Translate business use cases into reliable, scalable agent workflows with clear handoffs and fallbacks
- LLM & RAG Engineering
- Develop and optimize prompt engineering strategies for reasoning, task execution, and tool use.
- Implement Retrieval‑Augmented Generation (RAG) pipelines using structured and unstructured data.
- Design and manage agent memory systems, including:
- Short‑term (conversation/state)
- Long‑term (vector‑based memory)
- Work with vector databases
- Tooling, APIs & Integration
- Build and integrate internal and external tools (APIs, microservices, enterprise systems).
- Enable agents to safely interact with:
-
- Databases
- Business applications
- Automation platforms
- Cloud services
- Ensure robust error handling, retries, and fallback logic.
- Production Readiness & Operations
- Deploy agentic systems into production environments.
- Implement monitoring, logging, and evaluation frameworks for:
- Accuracy and task success
- Latency and cost efficiency
- Safety and policy compliance
- Design and enforce guardrails to prevent hallucinations, unsafe actions, or data leakage.
- Continuously optimize system performance and cost
- Collaboration & Delivery
- Partner closely with business stakeholders, application developers, data teams, and business stakeholders.
- Act as a technical product owner for agentic solutions: collect user and business requirements, clarify problem statements, and translate them into system designs and implementation plans
- Identify high‑impact AI opportunities and rapidly prototype, validate, and scale solutions.
- Contribute to internal best practices, architecture standards, and AI governance.
- Lead agentic tooling decisions: evaluate frameworks, model providers, orchestration patterns, and integration approaches; make pragmatic build/buy choices aligned to security, reliability, and cost constraints.
- Act as a technical product partner: collect user requirements, define success metrics and acceptance criteria, translate needs into solution designs and implementation plans, and drive delivery from prototype to production.