Software Engineer, ML Infra

GTV•San Francisco, CA

1d•$225 - $275

About The Position

About Unit Unit is an AI video company. We’re building GTV, a consumer product that combines creativity and automation to deliver a next-generation video experience. Our team includes senior people from Instagram, TikTok, and NVIDIA. We’re based in San Francisco and backed by Khosla Ventures, Air Street Capital, and other leaders in AI. Role Overview We're looking for an experienced ML Infra Software Engineer to join the GTV team. You'll work closely with researchers and engineers to build our core ML systems, including training and serving generative models, orchestrating real-time data pipelines, enabling reinforcement learning at scale, and more. This role is for someone who loves building high-performance ML infra and wants to reimagine the future of media.

Requirements

5+ years of software engineering experience and proficiency in Python
Industry experience designing and operating large-scale distributed systems for ML training and serving
Deep understanding of multi-node training, reinforcement learning, large-scale data processing, and real-time feedback loops
Proven ability to debug complex systems, navigate ambiguity, and take ownership of challenging problems
Excellent communication and collaboration with cross-functional teams (research, engineering, and design)

Nice To Haves

Prior work experience at early-stage or high-growth startups
Familiarity with video generation, evaluation, and labeling
Background in generative AI and infra for consumer-facing products

Responsibilities

Design and develop scalable ML systems that power the creative engine behind GTV
Partner with researchers and engineers to turn new model/product ideas into production-ready workflows
Build robust pipelines for real-time data collection, labeling, and feedback
Drive reliability and efficiency of distributed systems with automated testing, performance benchmarking, and end-to-end observability