Ai Ml Engineer

Key Responsibilities
Agentic & LLM System Development
Build agentic AI systems, including agent orchestration, planning loops, tool calling, and memory modules.
Implement LLM toolchains, custom prompts, templates, evaluators, and multi-step reasoning workflows.
Develop autonomous/semi-autonomous agents for retrieval, summarization, decision support, or workflow automation.
RAG Pipelines & Vector Intelligence
Design and implement RAG pipelines end-to-end: ingestion, chunking, embeddings, indexing, vector search, hybrid retrieval, and grounding.
Integrate vector databases such as pgvector, Pinecone, Weaviate, or Milvus.
Optimize retrieval quality, latency, and factual accuracy using rerankers, retrieval evaluators, and freshness pipelines.
Model Integration & AI Ops
Integrate enterprise-grade AI APIs, foundation models, and transformer models into scalable systems.
Implement robust evaluation frameworks including offline and online evals, regression tests, content safety checks, and red-team scenarios.
Build monitoring for model drift, agent failure modes, hallucination detection, and end-to-end system health.
Full-Stack & Cloud Alignment
Deploy services using Python/Node/Java microservices, serverless functions, or containerized workloads.
Integrate event streams, API gateways, and cloud-native patterns across Azure/AWS/GCP.
Build CI/CD pipelines for AI services with safe rollouts, versioning, and feature flags for model updates.
Prototyping & Rapid Iteration
Rapidly experiment with models, embeddings, architectures, and agentic patterns.
Translate business requirements into AI-native technical architectures and communicate trade-offs via demos and deep-dives.
Document designs, experiments, and evaluation results for reproducibility and knowledge sharing.
Professional & Technical Skills
Must-Have (AI Native Core)
Hands-on experience with LLMs, agent frameworks (LangChain, LlamaIndex, Semantic Kernel, LangGraph, AutoGen), and vector DBs.
Strong Python skills for building agents, RAG services, data pipelines, and evaluation harnesses.
Deep understanding of transformer models, embeddings, prompt engineering, and model fine-tuning workflows.
Experience deploying AI systems to production with monitoring, error handling, retries, and fallback strategies.
Good to Have
Experience with cloud AI platforms (Azure OpenAI, Bedrock, Vertex AI).
Knowledge of multimodal models, reinforcement learning, or advanced reasoning agents.
Familiarity with evaluation frameworks (pytest, JUnit) and custom eval harnesses.