AI Engineer
UMATRAI Engineer
We are looking for an AI Engineer with experience building AI-powered product features using modern foundation models. This role focuses on applied AI, prompt engineering, system orchestration, and product integration, not core model training or ML infrastructure.
The ideal candidate deeply understands LLM behaviour, strengths, and failure modes and knows how to turn models into reliable, scalable, user-facing capabilities.
What You Will Do
* Design and implement LLM-powered product features using prompt engineering, structured outputs, and tool or function calling
* Build prompt pipelines for complex tasks such as reasoning, extraction, classification, summarization, and planning
* Develop and maintain retrieval-augmented generation systems, including chunking, embeddings, retrieval strategies, and prompt composition
* Evaluate and iterate on prompt quality using both qualitative and quantitative methods
* Orchestrate multi-step AI workflows including agents, tools, memory, retries, and fallbacks
* Integrate AI systems into production services and APIs
* Collaborate with product and design teams to define AI-first user experiences
* Stay current with rapid developments in LLM capabilities, best practices, and emerging tools
What We Are Looking For
* 3+ plus years of experience building AI-driven or ML-adjacent product features
* Hands-on experience with LLMs including OpenAI, Anthropic, or open-source models and LLMOps
* Demonstrated expertise in prompt engineering
* Experience building retrieval-augmented generation systems
* Strong Python skills and experience integrating AI APIs into backend services
* Strong product intuition with the ability to balance quality, latency, and cost in real-world applications
* Familiarity with AI evaluation techniques including golden datasets, human review, and A/B testing
* Experience mitigating hallucinations, prompt injection, and safety issues
* Experience deploying custom models in production
* Must be based in the Bay Area or willing to relocate
Nice to Have
* Experience building agentic systems with tools, planners, and memory
* Familiarity with vector databases and embeddings
* Light experience with fine-tuning or adapters