AI Engineer

Build and ship AI systems

• Design, build, and operate LLM-powered agents and multi-step workflows — tool use, planning, state and memory, human-in-the-loop checkpoints, and sensible recovery when a step fails.
• Build forward-engineering agents: systems that take an ambiguous customer or business input and produce working technical output — queries, configurations, enrichment logic, briefs — that internal teams and customers then rely on.
• Take solutions into production and keep them healthy — deploy, monitor, and optimize for reliability and cost together.

Get the quality right

• Do prompt engineering properly: task decomposition, context design, structured outputs, tools and schemas, and the right model for each step. We treat prompts as engineering artifacts — versioned and reviewed like code.
• Build evaluation into the work — golden datasets, regression suites, and the metrics that matter for each workflow: accuracy, groundedness, hallucination rate, latency, and cost per unit of work.
• Know when an LLM is not the answer, and when retrieval, a classical model, or plain deterministic code is better.

Lift the team

• Mentor early-career engineers through code review, pairing, and design discussion. Most of the team is within a year of graduating and learns fast when it is taught well.
• Help set the standards the team works to — testing, prompt and evaluation conventions, observability, and what “done” means for an AI feature.
  • Work with product managers, platform engineers, and the teams who use what we build, across several time zones — translating fuzzy requirements into a technical approach with the trade-offs stated plainly. ​​​​​​​