We're looking for a data-obsessed intern to build evaluation systems for our AI-powered features - Hi Agent (our AI assistant that helps users build resumes and apply to offers) and Hi Scraper (our intelligent scraping workflow that extracts job offers from across the web).

You'll own the creation of both online and offline eval pipelines that allow us to measure and improve performance, experiment with different models and prompts, and ship improvements confidently without compromising quality.

What You'll Do

• Build baseline metrics and eval pipelines for Hi Agent and Hi Scraper
• Evaluate understanding, data extraction accuracy, tool choice and usage
• Create systems to safely experiment with different models and prompts
• Analyze real production data and identify improvement opportunities
• Work with existing tools and build upon them iteratively

Requirements

Must Have:

• Data obsession - comfortable spending hours analyzing rows of data
• Strong experience with LLMs and understanding of their behavior
• Analytical thinking and experimentation mindset
• Understanding of AI/LLM limitations and failure modes
• TypeScript proficiency (or willingness to deploy in TS while analyzing in Python)

Nice to Have:

• Familiarity with AI SDK or Mastra
• Knowledge of observability platforms, especially Langfuse
• Background in data analysis or statistics

What You'll Gain

• Own the eval infrastructure for production AI features
• Deep hands-on experience with LLM evaluation and improvement
• Space to experiment and make technical decisions
• High-velocity environment where you ship real improvements

Our Culture

Startup vibe meets high technical standards. We move fast and believe Done > Perfect. Attention to detail matters, but so does shipping quickly. Self-organized work style - you own your objectives and build.

Learn More

For a solid introduction to evals: https://www.youtube.com/watch?v=OHOZ5PgPj5M