Introduction to offer*
We need engineers who build AI systems that work outside of a Jupyter notebook. You’ll take architectures designed for real problems and turn them into production services — reliable, observable, and cost-efficient. Your day-to-day will involve writing Python, wiring up cloud infrastructure, and solving the unglamorous problems that make AI actually useful: data quality, latency, evaluation, and deployment.
You’ll work across the generative and classical AI stack — building knowledge-grounded AI systems, integrating LLMs into applications, training and deploying traditional ML models, and keeping it all running in production on Azure, AWS, or GCP.
Your tasks
Build and deploy knowledge-grounded AI systems end-to-end: data ingestion, chunking, embedding pipelines, retrieval logic, re-ranking, and response generation
Develop agentic applications — tool integrations, planning loops, memory management, guardrails — using frameworks like LangGraph, LangChain, Semantic Kernel, or equivalent
Implement and maintain ML pipelines for classical use cases: prediction, classification, recommendation, and optimization models
Deploy and optimize model serving infrastructure: API endpoints, batching, caching, GPU utilization, and cost management across cloud environments
Write clean, tested, production-grade Python — not prototype code that someone else has to rewrite
Build evaluation and monitoring pipelines: automated quality checks, drift detection, latency trac