Member of Technical Staff - Research Software Engineer - Safety Evaluations Infrastructure

Reflection AI
New York, San Francisco, London
Workplace: OnsiteFull timeFunction: Software EngineeringSkills: ["High ownership","Cross-functional collaboration","Discretion","Integrity","Bias toward action"]

Build and own secure, sandboxed infrastructure that powers sensitive safety evaluations, including CBRN and child-safety use cases. Design controlled data pipelines with least-privilege access, encryption, audit logging, and data-minimization safeguards. Create evaluation orchestration tooling for high-throughput runs in isolated environments, and integrate measurement results into systems that inform release decisions—while continuously improving security, reliability, and developer experience.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Reflection AI
Reflection AI
1 month ago

Member of Technical Staff - Research Software Engineer - Safety Evaluations Infrastructure

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 14 hours agoStatus: Live

Job Summary

Build and own secure, sandboxed infrastructure that powers sensitive safety evaluations, including CBRN and child-safety use cases. Design controlled data pipelines with least-privilege access, encryption, audit logging, and data-minimization safeguards. Create evaluation orchestration tooling for high-throughput runs in isolated environments, and integrate measurement results into systems that inform release decisions—while continuously improving security, reliability, and developer experience.
Location: New York, San Francisco, London
Workplace: Onsite
Employment Type: Full time
Job Function: Software Engineering
Seniority: Mid level

Key Responsibilities

  • •Design and build secure, sandboxed infrastructure for running sensitive model evaluations, including CBRN and other dangerous-capability domains.
  • •Build controlled data pipelines and storage for sensitive evaluation material with least-privilege access, RBAC, encryption, audit logging, and data-minimization safeguards.
  • •Partner with safety researchers and domain experts to translate evaluation designs into reliable, reproducible, and scalable systems.
  • •Build eval-orchestration tooling and harnesses for high-throughput evaluations in isolated environments.
  • •Implement guardrails, monitoring, and compartmentalization to keep sensitive work appropriately siloed across compute, data, and tooling.

Pay and Benefits

Perks:EquityHealth InsuranceDentalVisionLife InsuranceParental LeavePaid LeaveWellness StipendMeal AllowanceRemote Work

Key Requirements

  • •Strong software engineering skills, particularly in Python, with a track record building reliable, scalable infrastructure or platform systems.
  • •Experience building sandboxed/isolated security-sensitive execution environments (e.g., containerization, VM isolation, secure compute).
  • •Grounding in security engineering fundamentals: least privilege, need-to-know access, RBAC, secrets management, encryption, audit logging, compartmentalization, and defense-in-depth.
  • •Experience building data pipelines and handling sensitive or restricted data with appropriate safeguards.
  • •Ability to own problems end-to-end, including ambiguous, cross-functional ones.
Experience:Trust and SafetyML evaluationLLMsAgentsDangerous-capabilityDual-use
Skills:High ownershipCross-functional collaborationDiscretionIntegrityBias toward action
Tech Stack:PythonContainerizationVM isolationSecure computeRBACSecrets managementEncryptionAudit loggingData-minimizationSandboxed executionVMLLMsAgents

Company Brief

Reflection AI
Builds frontier autonomous AI systems focused on autonomous coding agents (product: Asimov) to create organizational superintelligence, founded by former DeepMind/Google researchers and hiring across SF, NYC, London.
Industry: AI & Machine Learning
Company Size: Small (11 to 50 employees)
Growth: Growth Stage Startup
Funding: Series A
Headquarters: San Francisco, United States
Founded: 2024
WebsiteLinkedInGlassdoor