AI Safety Researcher
Probe model behaviour for failure modes and build the evaluations that gate releases across Fivetran's data platform.
1638 roles across 160 companies — filter by region or category.
Showing 50 of 1,092 matching positions (201–250)
Clear filtersProbe model behaviour for failure modes and build the evaluations that gate releases across Fivetran's data platform.
Turn model capabilities into dependable product surfaces across Fivetran's data platform, balancing latency, cost and quality.
Build the internal platform that lets product teams ship to Fly.io's developer platform safely and often.
Train, evaluate and serve models that power Fly.io's developer platform at production scale.
Own the lakehouse and streaming infrastructure underneath Fly.io's developer platform, with an eye on cost and freshness.
Own the transformation layer and semantic definitions behind reporting on FusionAuth's security platform.
Build detections and response automation covering the infrastructure behind FusionAuth's security platform.
Own the modelling loop behind GitHub's developer platform — features, training pipelines, offline evaluation and online experiments.
Design red-teaming and evaluation methodology for the models behind GitHub's developer platform, and turn results into mitigations.
Sit with customers, build the integrations they need on GitHub's developer platform, and feed the rough edges back to product.
Own Kubernetes, deployment tooling and paved-road abstractions underneath GitHub's developer platform.
Own the design system and the polish layer across GitLab's developer platform.
Probe model behaviour for failure modes and build the evaluations that gate releases across GitLab's developer platform.
Build detections and response automation covering the infrastructure behind GitLab's developer platform.
Own infrastructure-as-code, networking and capacity planning for GitLab's developer platform.
Investigate intrusions and turn each one into durable detection coverage for Glean's AI products.
Investigate training, alignment and evaluation methods that make the models behind Glean's AI products more useful.
Sit with customers, build the integrations they need on Glean's AI products, and feed the rough edges back to product.
Investigate intrusions and turn each one into durable detection coverage for Grafana Labs's developer platform.
Build the ingestion, storage and query layers that every team uses to reason about Grafana Labs's developer platform.
Deliver working implementations of Grafana Labs's developer platform inside customer environments.
Prototype and ship interface work for Grammarly's product suite, living between design and front-end code.
Own the AI surfaces inside Grammarly's product suite, deciding what to automate, what to assist and what to leave alone.
Sit with customers, build the integrations they need on Grammarly's product suite, and feed the rough edges back to product.
Run the large multi-team efforts behind Grammarly's product suite from kickoff through launch.
Make the build, test and deploy loop fast for everyone working on Groq's AI products.
Own infrastructure-as-code, networking and capacity planning for Groq's AI products.
Model warehouse data into tested, documented datasets the whole company trusts when analysing Groq's AI products.
Own Kubernetes, deployment tooling and paved-road abstractions underneath Gusto's employment platform.
Own the transformation layer and semantic definitions behind reporting on Gusto's employment platform.
Turn model capabilities into dependable product surfaces across Gusto's employment platform, balancing latency, cost and quality.
Build the internal platform that lets product teams ship to Gymdesk's product suite safely and often.
Build agentic workflows and tool-calling systems on top of Gymdesk's product suite, and own their accuracy in production.
Partner with engineering teams on threat models and secure defaults throughout Harvey's AI products.
Own the lakehouse and streaming infrastructure underneath Harvey's AI products, with an eye on cost and freshness.
Prototype and ship interface work for Harvey's AI products, living between design and front-end code.
Define evaluation criteria, pricing and rollout strategy for AI features across HashiCorp's developer platform.
Model warehouse data into tested, documented datasets the whole company trusts when analysing HashiCorp's developer platform.
Deliver working implementations of HashiCorp's developer platform inside customer environments.
Drive cross-team programmes across Hightouch's product suite, keeping dependencies, risk and timelines visible.
Build the pipelines, feature stores and monitoring that keep models behind Hightouch's product suite reproducible and observable.
Model warehouse data into tested, documented datasets the whole company trusts when analysing Hightouch's product suite.
Build the internal platform that lets product teams ship to HubSpot's product suite safely and often.
Review designs, harden services and fix classes of vulnerability across HubSpot's product suite.
Define evaluation criteria, pricing and rollout strategy for AI features across HubSpot's product suite.
Build the ingestion, storage and query layers that every team uses to reason about HubSpot's product suite.
Deliver working implementations of Hugging Face's open-source projects inside customer environments.
Own developer tooling, CI throughput and local environments for the teams building Hugging Face's open-source projects.
Build agentic workflows and tool-calling systems on top of Hugging Face's open-source projects, and own their accuracy in production.
Build the internal platform that lets product teams ship to Instacart's commerce platform safely and often.
No jobs match your search and filters. Try a different keyword or clear the filters.