Zoox dezvoltă tehnologii pentru mobilitate autonomă și o flotă creată special pentru servicii de mobilitate ca serviciu. Activitatea companiei reunește robotica, inteligența artificială, ingineria auto, vehiculele electrice și sistemele avansate de senzori pentru a crea soluții de transport mai sigure, mai curate și mai plăcute. Compania își propune să redefinească modul în care oamenii experimentează mobilitatea personală prin vehicule complet autonome, proiectate în funcție de nevoile pasagerilor.
Contract Student AI Validation and Benchmarking Engineer
Evaluate AI agent accuracy and build datasets for benchmarking. Support Zoox’s autonomous robotaxi technology with Python tools and analysis.
Descriere
- Contribute to the end-to-end validation pipeline for AI tools.
- Maintain test datasets.
- Operate benchmark pipelines and analyze results to spot routing errors and regressions across agent variants.
- Create and expand ground truth datasets to compare agent outputs with verified answers.
- Find gaps in benchmark validation and help strengthen evaluation infrastructure.
- Define evaluation measures, including label accuracy and structured output correctness.
- Investigate agent output failures and work with engineers to identify practical improvements.
- Build scripts and tools to automate data collection, result parsing, and metric reporting.
- Record findings, monitor benchmark trends, and share results with the team.
- Collaborate with engineers and researchers developing autonomous technology.
Cerințe
- Be currently enrolled in a B.S. or M.S. program in computer science, data science, engineering, or a related discipline.
- Be available for an assignment of at least three months.
- Be able to work at least 20 hours per week.
- Be able to work on-site at one of the company’s offices.
- Follow Zoox confidentiality rules, including not using or sharing proprietary company information outside Zoox for academic research, theses, publications, or presentations.
- Have familiarity with Cursor or Claude.
- Have familiarity with Python.
- Understand evaluation measures such as precision, recall, F1 score, and confusion matrices.
- Be comfortable handling structured data formats such as CSV and JSON.
- Have experience writing or modifying reproducible analysis scripts.
- Have prior exposure to LLM-based systems, prompt engineering, or AI agent evaluation.
- Have experience with Jira or Slack, such as ticketing or messaging tools.
Beneficii
- Ongoing, meaningful work that can be done alongside academic studies.
- Chances to learn and build new skills.
- Practical experience in a professional setting.
- Benefits eligibility is determined by the vendor.