RL Environment Engineer, Verita AI

ML Engineer

$104-312k

Python

CUDA

PyTorch

Mid and Senior level

San Francisco Bay Area

Remote from Canada, Europe, UK, US

High-trust data pipelines for training and evaluating models

Be an early applicant

High-trust data pipelines for training and evaluating models

1-20 employees

B2BArtificial IntelligenceMachine LearningSaaS

Be an early applicant

$104-312k

Python

CUDA

PyTorch

Mid and Senior level

San Francisco Bay Area

Remote from Canada, Europe, UK, US

1-20 employees

B2BArtificial IntelligenceMachine LearningSaaS

Company mission

To bring human taste, safety, and rigor into the next generation of intelligent systems.

Job

Company

Role

Who you are

Master’s degree in Computer Science, AI, ML, or a related technical field
Strong Python skills with production-quality engineering standards
Experience designing or working with RL environments or training pipelines
Solid understanding of modern LLMs and their limitations
Ability to work quickly, iterate reliably, and respond to feedback
Advanced English proficiency (C1/C2)

Desirable

Deep knowledge of transformer internals or LLM training/inference
Experience with inference libraries such as vLLM or SGLang
CUDA or custom kernel optimization experience (e.g. Pallas)
Research experience with publications or high-quality open-source work
Experience building complex or open-ended RL-based learning systems

What the job involves

Design and build reinforcement learning environments for training and evaluating LLMs
Translate modern ML and AI research into structured RL problems
Implement reliable, debuggable, and scalable training environments in Python
Collaborate with researchers and engineers to improve model learning quality
Complete an average of two well-scoped tasks per week
Iterate quickly based on feedback and evaluation results

Application process

Submit application
Complete a technical take-home assignment
Review and discussion with the Verita AI team

Our take

Verita AI is building the human data layer for the next generation of artificial intelligence. So instead of focusing purely on raw code or algorithms, Verita helps AI systems learn in ways that feel human.

The company offers curated datasets and evaluations that give machine learning models the quality training you'd normally see in craft-focused creative work. It also partners closely with teams to co-design data pipelines and fine-tune processes to make AI products reliable, safe and ready for real-world use.

While the company may be young, it's onto something as the world leans harder into multimodal AI. With founders and backing by experts from places like Stanford, as well as support from leading AI researchers, Verita AI is betting on the idea that the future of intelligence isn't just about crunching facts but on capturing human taste too.

Freddie

Company Specialist at Welcome to the Jungle

Company

Leadership

Rithika K.

(Co-Founder & CEO)

Prior to co-founding Verita AI in 2025, Rithika was part of the Executive Ops Team at Mercor.

Rishi K.

(Co-Founder & CTO)

Previously, Rishi was a Member of Technical Staff at Oracle and Broadcom.

Salary benchmarks

We don't have enough data yet to provide salary benchmarks for this role.

Submit your salary to help other candidates with crowdsourced salary estimates.

Share this job