Member of Technical Staff, AI Data and Evals Engineer at Tech Careers — San Francisco, CA
Зарплата не указана
ОткликнутьсяКомпания: ••••••••
Локация: San Francisco, CA
Формат: не указан
Занятость: Полная занятость
Грейд: Senior
mggk8RM is sharing this role on behalf of Tech Careers.
On-site (San Francisco Bay Area, CA) | $200K to $400K | 5 openings PLEASE INCLUDE YOUR LINKEDIN TO BE CONSIDERED!
About the company
An early-stage, venture-backed AI company that helps enterprises in engineering and life sciences put AI to work on their highest-value workflows, then turns that real professional work into training and evaluation data for frontier AI labs.
Enterprise projects are the core of the business, and the lab data work compounds on top of them, which sets it apart from expert networks and pure data vendors.
The team is 11 people today, founded by operators from a fast-growing AI data company and top quantitative trading firms, with more than a dozen hires planned. The role This is the customer-facing engineering seat on the team.
You'll partner directly with research teams at AI labs to decide what a training or evaluation dataset needs to measure, design how it gets built, and confirm the finished data actually does what the researchers asked for.
Day to day you'll write code alongside domain experts who convert enterprise data and expert workflows into model tasks, and you'll spend real time reading examples, digging into where models fail, and tightening how quality gets measured.
The team is hiring 5 engineers into this track at once, so you'll be shaping the playbook with a new cohort rather than filling a single seat. Goals are set at a high level and you're trusted to get there without being micromanaged. What you'll do
• Scope dataset requirements with lab researchers and review early samples with them before a dataset scales to full production
• Build the pipelines and agent workflows that generate tasks, produce reference answers, and validate data before delivery
• Design quality checks and grading approaches with domain experts in engineering and life sciences, catching unclear instructions, wrong answers, and tasks a model could pass for the wrong reasons
• Run down problems surfaced in review or after delivery, fix the underlying cause, and carry the lessons into the next dataset
• Stay with each dataset after handoff, taking researcher feedback on how it performs and turning it into concrete changes
• Work with enterprise customers on the deployments that supply the underlying professional workflows
What you must bring
• Strong hands-on programming ability, shown by shipping products, agents, or workflow automation
• Solid backend or full-stack skills in Python and TypeScript
• Experience building data pipelines, automation, or AI applications
• Real experience with at least one of model evaluation, synthetic data generation, or agent systems
• The ability to take an open research question, turn it into a technical plan, and judge honestly whether it worked
• A high bar for data quality, including the patience to read individual examples and trace inconsistencies to their source
• Clear, direct communication with researchers and subject matter experts, especially while requirements are still taking shape
• Comfort working face to face with AI lab and enterprise customers
• No minimum years of experience. Technical ability matters more here than tenure
Nice to have
• Prior work at a data pipeline or AI data business
• Experience at an early-stage startup or on enterprise deployments
• Experience working independently toward loosely defined goals
Compensation & logistics
• $200K to $400K base salary plus competitive equity and commission
• On-site five days a week in the San Francisco Bay Area, CA (relocation support is not offered)
• 5 openings
• Open to visa sponsorship, including new H-1B and TN, as well as visa transfers such as OPT and H-1B transfers
This isn't the right fit if
• Your strengths are mainly customer-facing and you don't write production code regularly
• You'd rather not inspect individual data examples by hand
• You work best with close day-to-day direction rather than high-level goals
• You can't work on-site in the Bay Area five days a week