AI Agent Engineer

Benchmark Sports; Irvine (CA); On-site; USD 100,000 - 140,000; Full time

Benefits offered by this job

Full health + dental coverage

401(k)

Daily team lunches

Weekly team events (including golf)

Job summary

A tech sports company located in California is seeking experienced engineers to build AI coaching agents for real-time athlete feedback. The role demands strong foundational knowledge in machine learning and full stack experience while working closely with a small, talented team. Ideal candidates will thrive in a fast-paced environment, shipping innovative product features that enhance training experiences. Benefits include full health coverage, 401(k), and regular team events.

Qualifications

• Experience in building AI coaching agents for real-time feedback.
• In-depth knowledge of LLMs and coding workflows.
• Experience in shipping products that focus on user experience.

Responsibilities

• Build AI coaching agents that give athletes real-time feedback on their swing.
• Own the LLM stack: prompts, tool use, RAG, context management.
• Ship multi-turn coaching conversations that hold context across sessions.

Skills

Strong ML fundamentals

Full stack experience with good taste for UI/UX

Up to date with latest agentic coding workflows

Experience with LLMs

Job description

Why Benchmark Sports

Benchmark Sports is redefining how golf is learned through real‑time simulation and AI coaching. We’re profitable, moving fast, and scaling globally. Joining now means owning core systems used by millions of athletes. Shipping product that changes how people train.

The Culture

High ownership, touching all elements of the product full stack, owning business outcomes on a small lean team of highly technical people. We have a high talent bar and only hire people capable of working full stack and having a clear understanding of the product mission. In office in Irvine. Work 6 days a week.

What You’ll Work On

• Build AI coaching agents that give athletes real‑time feedback on their swing
• Own the LLM stack: prompts, tool use, RAG, context management
• Ship multi‑turn coaching conversations that hold context across sessions
• Build evaluation and grounding systems that catch hallucinations against real swing data
• Get response latency low enough for live coaching

Core Expectations

• Strong ML fundamentals
• Up to date with latest agentic coding workflows
• Full stack experience with good taste for UI/UX
• Worked on either cutting edge implementations of LLMs or shipped LLM products used by real users

Who Thrives Here

Young, ambitious builders with a blank‑slate mindset. Founder energy or campus incubator DNA. Someone who has shipped things end‑to‑end and owns the outcome. Strong preference for candidates from top technical environments (Stanford, Berkeley, Cal Poly, CMU), but we’ll hire the best builder regardless of pedigree. If you need a roadmap to do your best work, this role will be a bad fit.

What Success Looks Like: First 90 Days

• Ship a coaching agent that gives athletes useful feedback on at least one swing dimension
• Set up automated metrics for coaching accuracy, hallucination rate, and response quality
• Get response latency into the real‑time window
• Know where the agent breaks and have a plan to fix it
• Release to several sports beyond golf: padel, tennis, cricket, etc.

Benefits

• Full health + dental coverage
• 401(k)
• Daily team lunches
• Weekly team events (including golf)