/draftSign in

AI Engineer — Learn Engine: Intelligence & Optimization

Hellyeah AI · Hybrid · San Francisco HQ · $200k–$1000k

Posted by Hellyeah AI on their careers site · View the original posting

Draft this job a résumé tailored to it, and the application filled in

draft's reading

Open 88 days

The posting, as Hellyeah AI wrote it

Build the brain of an autonomous growth OS. The system you create will manage millions in ad spend and get measurably smarter with every dollar. This is the moat — every competitor has humans optimizing campaigns manually. You are building the intelligence layer that compounds. The Platform engineer creates the tools, you create the decisions. Together you build something nobody else has.

Must Have

Has built recommendation, optimization, or decision systems where outputs improve future inputs.

  • Strong statistical reasoning and experimentation judgment under noisy real-world data.
  • Strong LLM orchestration or agent-system experience for reasoning over campaign context.
  • Can design optimization policies, scoring systems, or automated recommendation loops.
  • AI-first development workflow and ability to ship production systems quickly.

Nice to Have

Ad-tech optimization patterns (bid management, budget allocation, ROAS optimization)

  • Reinforcement learning (RL) experience is a plus
  • Hyperparameter optimization (HPO) experience is a plus
  • Model fine-tuning experience is a plus
  • Experience building agent-driven automation (LLM agents that take actions)
  • Background in growth engineering, performance marketing, or data science
  • Experience with Mastra or similar agent orchestration framework

Own the intelligence and optimization layer of Learn Engine.
Build recommendation engines for bid changes, budget reallocation, pause/boost decisions, and postback optimization.
Turn SSOT campaign data into high-quality optimization guidance and closed-loop decision systems.
Define how the system learns from outcomes and continuously improves campaign strategy over time.
This role owns decision quality, optimization policy, and learning loops — not platform plumbing or simulator infrastructure.