Dataset Lead, Software Engineering

Reinforced Learning; Town of Tonawanda (NY); On-site; USD 150,000 - 210,000; Full time

Job summary

Reinforced Learning in the United States seeks a senior software engineer who will own and scale the pipeline turning experienced engineers into demonstrations, reviews and preference data. You will lead program scope with research partners and guide a contributor bench of senior engineers.

You bring 5+ years of software engineering, leadership or code reviews, and a proven ability to articulate engineering quality.

Qualifications

• Five or more years as a software engineer with leadership or review experience.
• Strong opinions about engineering quality and ability to document them.
• Experience running a scalable process: hiring, onboarding, QA, feedback.
• Comfort reading trajectories and diagnosing what went wrong.

Responsibilities

• Own the pipeline turning engineers into demonstrations, reviews and preference data.
• Scope coding programs with partner research teams: languages, task types, trajectory formats, review depth.
• Lead a contributor bench of senior engineers, as contributors and reviewers.
• Define a good trajectory: reasoning, tool use, testing, communication.
• Calibration sessions so reviewers grade the same work the same way.
• Quality and throughput tracking, with early intervention when either slips.

Skills

Software engineering

Leadership or review

Engineering quality advocacy

Documentation of opinions

Scaled processes (hiring, onboarding,

Trajectory analysis

Tools

Code review tooling

CI systems

Developer productivity tooling

Job description

The role

Our largest programs are in code. You’ll own the pipeline that turns experienced engineers into demonstrations, reviews and preference data of a standard a model should learn from, and keep that pipeline honest as it scales.

What you’ll own

• Scoping coding programs with partner research teams: languages, task types, trajectory formats, review depth.
• A contributor bench of senior engineers, as both contributors and reviewers.
• The definition of a good trajectory: reasoning, tool use, testing, communication.
• Calibration sessions so reviewers grade the same work the same way.
• Quality and throughput tracking, with early intervention when either slips.

What you bring

• Five or more years as a software engineer, with time spent leading or reviewing others.
• Strong opinions about engineering quality and the ability to write them down.
• Experience running a process at scale: hiring, onboarding, QA, feedback.
• Comfort reading trajectories and diagnosing what went wrong.

Helpful, not required

• Work on code review tooling, CI systems or developer productivity.
• Prior work on training data or evaluation for coding models.