Remote AI Safety & Trust Product Manager

Innodata Inc.; United States; Remote; USD 180,000 - 220,000; Full time

Job summary

Innodata Inc. is seeking a Technical Product Manager for Trust & Safety to define and ship data products used to train and evaluate models for safe personalization, content moderation, and related tasks. The role focuses on building RL environments with traps and ensuring safety rubrics guide model behavior.

You will own end-to-end product lifecycle, collaborating with ML researchers, engineers, red-teamers, and policy experts to align product outcomes with safety goals.

Qualifications

• 5+ years of product management experience in ML/AI data, security, or trust & safety.
• Hands-on familiarity with how LLMs are trained and evaluated: SFT, RLHF/RLAIF, reward models, verifiers, benchmarks.
• Working knowledge of agentic AI: tool use, multi-step planning, and associated failure modes.
• Strong AI safety understanding: reward hacking, deceptive alignment, jailbreaks, prompt injection.
• Experience designing metrics that capture risk, system performance, and user impact.

Responsibilities

• Define data products for frontier labs, translating research goals into clear specs for safety training data, RL environments, and evaluations.
• Design trap-based RL environments with misalignment and cyber traps to test model behavior.
• Specify safe-behavior rubrics, verifiers, and reward signals to guide model learning.
• Define measurable safety criteria for each product, including attack success, over-refusal, trap trigger rate, and performance under adversarial pressure.
• Collaborate with researchers, ML engineers, red-teamers, policy experts, and annotation leads to deliver safe data products.

Skills

Product management

ML / AI data

Trust & Safety

RLHF/RLAIF

Reward models

Safety evaluation

Stakeholder communication

Job description

Innodata Inc. is seeking a Technical Product Manager for Trust & Safety to define and ship data products used to train and evaluate models for safe personalization, content moderation, and related tasks. The role focuses on building RL environments with traps and ensuring safety rubrics guide model behavior.

You will own end-to-end product lifecycle, collaborating with ML researchers, engineers, red-teamers, and policy experts to align product outcomes with safety goals.