Posted 23 August, 2026
AI Product Experience Compliance
HCLTech
201301, UP, IN
Full Time
Reference: 1324a321fcf29a9f
Job Description
Role- AI Product Experience Compliance
\nLocation- Noida & Hyderabad
\nExperience- 14-21 Years
\nRole Summary:
\nDefine the AI harnessing and compliance specification — safety, guardrails, explainability, auditability — and build the validation mechanism the CTO org's builds are tested against. Secondary but connected scope: identify where user-experience friction (e.g., excessive steps to complete a task, unclear agentic responses) creates compliance or safety risk — specifically where friction causes users to bypass intended safeguards, not UX quality in general.
\nKey Responsibilities:
\nCore (compliance/safety):
\n- \n
- Specify AI guardrails: prompt/response policy, PII handling, refusal and escalation behavior \n
- Build automated compliance checks and red-team test suites into the validation harness \n
- Maintain the audit trail and evidence pack for internal and customer audits \n
- Track applicable AI regulation (EU AI Act, sector-specific rules) and translate into engineering requirements, in partnership with legal/compliance review — this role proposes interpretations, it does not have unilateral legal authority \n
- Validate delivered AI behavior against the harnessing spec at each release \n
- Evaluate model outputs for bias/fairness issues relevant to the applicable regulatory framework \n
Secondary (experience-risk signal):
\n- \n
- Capture specific UX-compliance-risk signals — e.g., number of steps/clicks to complete a task, points where users abandon or override a safety confirmation, ambiguity in agentic responses that leads to user error — and flag where these create compliance exposure, distinct from general UX quality feedback \n
Must-Have Requirements:
\n- \n
- 12+ years full-stack engineering with hands-on LLM application work \n
- Has implemented guardrails, content filtering, or model-output validation in a production system — be ready to describe the architecture and a specific failure it caught \n
- Deep understanding of prompt injection, jailbreak techniques, and data-leakage failure modes \n
- ML evaluation experience, including bias/fairness evaluation methodology \n
- Regulated-industry exposure (finance, healthcare, or similar) with direct experience operating under an external audit \n
- Hands-on experience with named red-teaming/guardrail/evaluation tooling — specify which from: Garak, PyRIT, Promptfoo, NeMo Guardrails, Llama Guard, Guardrails AI, TruLens, Ragas, or equivalent \n
- Can read regulatory text (EU AI Act or sector-specific) and produce draft test cases, with the explicit understanding that a compliance/legal function reviews the final interpretation \n
Preferred:
\n- \n
- UX research or product analytics background (task-completion metrics, friction analysis) — genuinely useful for the secondary responsibility, but should not be must-have alongside the security/compliance bar \n
- Experience with formal AI governance frameworks (NIST AI RMF, ISO/IEC 42001) \n
- Prior work directly supporting an external regulatory audit (not just internal compliance) \n