Technical Director - AI Roblox Studio

Workman Labs; San Mateo (CA); Hybrid; USD 399,000 - 458,000; Full time

Job summary

Workman Labs in San Mateo, CA, is seeking a hands-on director-level engineer to own the technical direction of agentic AI systems embedded in Roblox Studio as it evolves into an AI-Native IDE. You will shape how coding models, agents, and context management are used across the Assistant and creator workflows, building production-grade systems and evaluation pipelines.

This role demands deep systems expertise, ownership of large-scale platforms, and the ability to lead across multiple teams as an

Qualifications

• 12+ years of engineering experience, including ownership of large-scale, long-lived platforms.
• Hands-on experience building LLM-powered and agentic systems in production, beyond chat-based interfaces.
• Strong systems and software engineering skills — able to design the architecture and write the code.
• Track record defining evaluation frameworks for AI product quality, combining automated metrics with human review, and optimizing inference for latency, throughput, and cost.
• Strong background in developer tools or IDE-like platforms; proven ability to set technical direction across multiple teams as an individual contributor.
• Preferred: background in coding models or code-generation systems, and/or research experience in reinforcement learning (RLHF, RLAIF, preference modeling).

Responsibilities

• Design and build the agentic AI systems at the core of the AI-Native IDE (planning, execution, validation, evaluation) so they are production-grade and reliable.
• Decide how Studio uses coding models, agents, retrieval, tool calling, and context management across the Assistant and the broader creator workflow.
• Own the quality bar and evaluation strategy, standing up quantitative and qualitative eval pipelines including human-in-the-loop.
• Optimize systems for latency, throughput, and cost (caching, routing, distillation, quantization).
• Bring reusable patterns and best practices to teams across Studio, raising applied-AI fluency of Studio engineering.

Skills

Engineering leadership

LLM-powered systems

Performance optimization

Architectural design

Quantitative evaluation

Tools

Coding models

Code generation

Job description

Hands-on individual-contributor role owning the technical direction of the agentic AI systems embedded in Roblox Studio as it evolves into an AI-Native IDE where intelligent systems plan, act, validate, and iterate alongside creators.

$399,420 – $457,970 / yr

Location

San Mateo, CA

Hybrid

Seniority

Director

Compared with 41 other Engineering roles: Common across similar roles Unique to this role

What you'll do

Responsibilities

• Design and build the agentic AI systems at the core of the AI-Native IDE (planning, execution, validation, evaluation) so they are production-grade and reliable Common
• Decide how Studio uses coding models, agents, retrieval, tool calling, and context management across the Assistant and the broader creator workflow Common
• Own the quality bar and evaluation strategy, standing up quantitative and qualitative eval pipelines including human-in-the-loop
• Optimize systems for latency, throughput, and cost (caching, routing, distillation, quantization)
• Bring reusable patterns and best practices to teams across Studio, raising applied-AI fluency of Studio engineering

What you'll need

Experience

• 12+ years of engineering experience, including ownership of large-scale, long-lived platforms
• Deep, hands-on experience building LLM-powered and agentic systems in production, beyond chat-based interfaces Common
• Strong systems and software engineering skills — able to design the architecture and write the code
• Track record defining evaluation frameworks for AI product quality, combining automated metrics with human review, and optimizing inference for latency, throughput, and cost
• Strong background in developer tools or IDE-like platforms; proven ability to set technical direction across multiple teams as an individual contributor
• Preferred: background in coding models or code-generation systems, and/or research experience in reinforcement learning (RLHF, RLAIF, preference modeling)