About the role

Turing builds the training and evaluation data powering the world’s leading frontier AI labs. As these labs push the frontier, they bring us their newest models and ask the critical question: how does our model perform against your benchmark datasets, where does it fail, and what data will make it climb the hill?

As a Field Engineer, AI Labs, you are the technical face of Turing to frontier research teams. You sit at the exact intersection of client-facing solutions engineering and hands-on AI evaluation. When a frontier lab provides a new model endpoint, you wire it up, run comprehensive evals across Turing’s datasets, synthesize raw performance metrics into actionable intelligence, and present those findings directly to client researchers and technical leaders.

This is not a pure research engineering seat and not a traditional forward-deployed IT role. It is a high-velocity, high-agency solutions engineering role for a technically fluent, dependable builder who thrives on talking to customers, running benchmarks across model APIs, and packaging last-mile data deliveries with extreme precision.

Why this role matters

You will work directly with research scientists and technical executives at top AI labs (such as xAI, Anthropic, OpenAI, and emergent frontier labs). Every eval you run and every benchmark report you write directly influences how frontier labs evaluate model capabilities, uncover failure modes, and procure the pre-training and post-training data required for next-generation intelligence.

What you’ll own

Evaluation & Benchmarking

• Connect frontier model APIs and inference endpoints (OpenAI, Anthropic, AWS Bedrock, vLLM, private client gateways) into evaluation harnesses.
• Execute large-scale benchmark runs across proprietary Turing datasets (reasoning, coding, multimodal, agentic environments, and GDP Val).
• Analyze eval results, isolate failure cases, detect model behavioral drift, and identify the exact data distributions needed to address capability gaps.

Technical Pre-Sales & Client Engagement

• Serve as the technical counterpart to sales leads on calls with frontier AI labs and enterprise AI teams.
• Translate complex evaluation metrics into compelling, high-signal reports, client presentations, and technical collateral.
• Answer technical questionnaires and RFIs from lab researchers regarding data design, task curation, difficulty stratification, and rubrics.
• Surface lab feedback and emerging benchmark requirements back to internal R&D and dataset design teams.

Internal Data Systems & Catalog Stewardship

• Navigate and operate Turing’s internal data infrastructure (DataOS, Data Hub, and task catalogs) to query, manipulate, and move data pipelines.
• Slice and curate custom data subsets based on specific difficulty thresholds, benchmark rubrics, or top-performing data points.
• Audit data provenance and maintain up-to-date specifications across Turing's public catalog and internal dataset inventories.

Packaging & Last-Mile Delivery

• Own last-mile delivery, schema alignment, and packaging so that data delivered matches client technical specifications flawlessly.
• Work closely with engineering and quality leads to verify that all data deliverables adhere to rigorous validation standards before client handoff.

What you bring

Required

• 1 to 5 years of professional experience in data science, AI/ML engineering, solutions engineering, or data analytics with a demonstrated transition into generative AI.
• Strong hands-on Python proficiency: working with REST/gRPC model APIs, inference endpoints (OpenAI, Anthropic, Bedrock, vLLM), pandas, and SQL.
• Practical experience with model evaluation, benchmark design, or LLM-as-a-judge pipelines. You know how benchmarks are structured, scored, and debugged.
• High agency and a scrappy, startup mindset: you have owned projects end-to-end, delivered under tight deadlines, and thrive in fast-paced environments where playbooks are being built in real time.
• Exceptional written and verbal communication: ability to walk research scientists through technical evaluation metrics and command confidence on commercial calls.
• Comfort working with command-line tools, internal tooling, and data manipulation workflows.

Preferred

• Prior experience at an early-stage AI startup or high-growth technology company where you wore multiple hats across engineering and customer support.
• Hands-on experience with evaluation harnesses and agent benchmarks (e.g., Harbor, Terminal-Bench, SWE-bench, Inspect) or agentic workflows.
• Experience fine-tuning models (e.g., LoRA/QLoRA), working with open-weights models, or conducting multimodal evaluations.
• Familiarity with data quality control, task verification, or synthetic data generation pipelines.

How we work

• High ownership, low ego: You own the last mile from intake to client handoff. The outcome matters more than who gets credit.
• Technical fluency with commercial empathy: We don't hide behind jargon. We present clear, honest numbers to lab researchers and act as their trusted technical partners.
• Rigorous data integrity: Provenance, versioning, and reproducibility come first. What we benchmark is what the client gets.

Location & Work Model

• Location: United States. Strong preference for candidates based in the San Francisco Bay Area (for regular interaction with local AI labs and team hubs); remote within the US considered for exceptional candidates.
• Employment Type: Full-time standard employee.

Compensation & Benefits

The annual base salary range for this full-time position is $175,000 – $232,000 USD. In addition to base salary, this role is eligible for equity, performance incentives, and comprehensive company benefits.

Individual compensation is determined by job-related skills, experience, qualifications, and relevant business considerations.