The Role

You will own the production system behind Turing’s software-engineering data programs, turning complex research requirements into predictable delivery across quality, throughput, contributor performance, timelines, and cost.

These programs may involve supervised coding demonstrations, repository-level tasks, agentic trajectories, reinforcement-learning environments, benchmarks, code review, and rubric-based evaluations. They can require coordinating hundreds of distributed software engineers while responding quickly to changing research requirements.

This is an operations leadership role with a meaningful technical bar. You must be able to inspect code, understand tests, interrogate quality signals, and challenge a workflow or rubric when it is not producing the intended result. You will not be expected to act as the principal engineer for every program. Your primary responsibility is to build and operate the system that consistently produces high-quality technical work at scale.

What You’ll Own

1) Operational execution — own end-to-end delivery on every project you run

• Design and manage data pipelines from customer specification to final delivery, with full accountability for scope, timeline, and quality.
• Scope and stand up coding workstreams across supervised demonstrations, agentic trajectories, RL environments, benchmark construction, and rubric-graded evaluation.
• Diagnose bottlenecks in real time — re-sequence workflows, refine instructions, create incentive systems, and scale review processes to hit throughput targets.

2) Quality ownership — ensure world-class data integrity on every project

• Own quality control across the annotation lifecycle: set the bar, measure against it, and close the gap when it slips.
• Analyze datasets to identify trends, anomalies, and systematic errors — then fix the root cause, not just the symptom.
• Implement and continuously improve annotation, evaluation, and curation best practices.

3) Large-scale coordination — orchestrate the work of 100–1,000+ contributors

• Define the required contributor profile and partner with talent teams to source, assess, onboard, and ramp distributed software engineers.
• Own contributor training, performance management, reviewer capacity, incentives, and corrective actions.
• Build team-lead and reviewer structures that maintain execution standards across programs involving hundreds of contributors.

4) Customer relationships — be the face of Turing to the world’s leading AI labs

• Act as the primary point of contact for researchers and program managers at frontier AI labs providing clear reporting on progress, quality, risks and recovery actions.
• Translate research intent into a task specification, and push back when a spec will not produce the signal the researcher actually wants.
• Build the kind of long-term trust that converts a one-off project into a multi-year partnership — and identify expansion opportunities along the way.

5) Playbook building — codify what works so future SPLs scale faster than you did

• Use Python, SQL or other appropriate tools to automate quality sampling, defect classification, throughput analysis, and weekly reporting.
• Turn successful workflows into reusable playbooks, quality controls, evaluation assets, and contributor-management systems.
• Share lessons and mentor other SPLs so each program improves the operating system for the next one.

What We’re Looking For

• Background in software engineering, technical program management, consulting, finance, startups, or other operationally intense environments, with a proven track record of managing complex, multi-stakeholder projects.
• Strong analytical and communication abilities: you can spot a bottleneck in a noisy production environment, build a measurement plan, and communicate the fix to a demanding client in plain language.
• Customer-facing experience: comfortable working directly with high-profile clients, managing expectations, and building long-term relationships.
• Excited by gritty process optimization and large-scale execution — you thrive on making complex operations faster, cleaner, and more reliable.
• You can read and review code. You can follow a pull request in Python, TypeScript, Java, or Go, read a test suite, and judge a delivered task independently.
• Bonus: Experience with agentic evaluation harnesses, software engineering benchmarks, or RL environments; Experience managing large distributed contributor networks or marketplaces; Prior work at an AI data vendor or a frontier lab

What Success Looks Like

30 days: Complete technical and operational calibration, establish the program baseline, validate acceptance criteria and delivery controls, and independently lead a defined workstream.

90 days: Deliver predictable throughput and quality, improve at least one material operating metric, maintain a trusted risk and reporting cadence, and demonstrate that defects are being detected internally before customer delivery.

180 days: Run concurrent programs with stable quality and cost performance, convert successful workflows into reusable assets, contribute evidence that supports account expansion, and help another lead or team adopt the operating system you built.

Why Turing

• Work directly with the world’s leading AI labs at the cutting edge of post-training, evaluation, and agentic AI research.
• Real impact on the path to AGI: the data you deliver will directly influence how frontier models are trained and evaluated.
• High ownership and influence. You will shape how Turing delivers at scale, with direct visibility to senior leadership.
• Direct-to-research customers. You will spend your time partnering with the people building the future of AI, not coordinating with procurement.