Jobs at hud

HUD is accepting applications for exceptional candidates across research, engineering, go-to-market, and operations. Applicants must propose a specific contribution based on research into HUD and provide concrete evidence of exceptional ability and relevant experience. The role is full-time, based in San Francisco, Singapore, or remotely with substantial time-zone overlap. The hiring process includes two technical interviews and a two- to three-day work trial.

Build privacy and anonymization systems that detect and transform PII, credentials, secrets, and other sensitive information in data used for AI training. Develop detection methods combining rules, statistical models, classifiers, and LLMs; create production data-processing pipelines; evaluate privacy risk and retained utility; and ensure robustness against schema drift, unusual formats, leakage, and adversarial re-identification.

Own end-to-end technical data and evaluation programs for frontier AI labs, coordinating requirements, milestones, vendors, quality assurance, delivery, and retrospectives. Translate ambiguous technical needs into executable plans, improve data quality and production workflows, manage cross-functional dependencies, and partner with research, platform, and marketplace teams to inform product improvements.

Own HUD’s production infrastructure and backend platform systems, focusing on reliability, scalability, performance, cost, deployment safety, and developer experience. Build AWS infrastructure with Terraform and Kubernetes, improve CI/CD and observability, design resilient services and asynchronous systems, manage incidents, and develop internal automation and tooling. The role supports high-availability AI, marketplace, and data-heavy workloads.

Lead HUD's data quality strategy for RL training data: build QC systems, validate synthetic data at scale, design evaluation experiments, partner with researchers and vendors, convert research insights into production validation pipelines and tools, and mentor research engineers.

Build full-stack product surfaces, backend services, APIs, databases, dashboards, internal tools, vendor workflows, and observability for reinforcement learning data and evaluation systems. Develop interfaces for browsing environments, inspecting trajectories, reviewing quality, debugging failures, and understanding model behavior. Partner with research, operations, vendors, and go-to-market teams to translate ambiguous requirements into reliable products. The role also involves production debugging and cloud infrastructure collaboration.

Build scalable QC systems and validation pipelines for RL training data: define quality standards, design metrics and experiments, audit supplier datasets, partner with vendors to diagnose failures, and integrate learnings into infrastructure and vendor tooling.

Design, implement, and maintain domain-specific benchmarks and infrastructure to run agents/models. Define tasks with subject experts, develop metrics and analyses, validate benchmark relevance to real-world evals, and produce clear technical documentation and reports.

The GTM Engineer is responsible for managing HUD's outbound sales pipeline, optimizing conversion metrics, and developing GTM workflows, targeting technical buyers. The role requires technical aptitude and sales skills relevant to AI and ML domains.

Lead implementation and deployments for frontier AI labs and data vendors: triage technical requests, build tools and one-off pipelines, diagnose ambiguous problems across code/data/environments, coordinate cross-functionally, and turn recurring manual work into reusable processes.

Research Engineers will build QA systems for training data, define quality standards, audit datasets, and collaborate with data vendors to enhance data generation processes.