AI Ethics & Governance Intern

As an AI Ethics & Governance Intern at Cehpoint, you will bridge the gap between cutting-edge AI engineering and responsible deployment. You'll author governance policies, conduct bias and fairness audits on production models, and ensure our enterprise AI systems align with global regulatory standards—directly influencing how secure, trustworthy, and compliant our AI products are in the market.

• Develop and maintain AI governance documentation including model cards, datasheets, and risk assessment reports aligned with NIST AI RMF and ISO/IEC 42001.
• Conduct fairness and bias audits on classification, generative, and LLM-based models using quantitative metrics (demographic parity, equal opportunity, calibration) and qualitative review.
• Partner with the AI Engineering team to integrate explainability tools (SHAP, LIME, Captum) into model evaluation pipelines and document interpretability findings.
• Map Cehpoint's AI systems against relevant regulatory requirements (EU AI Act, DPDP Act 2023, OECD AI Principles) and produce compliance gap analysis reports.
• Design and execute red-teaming and adversarial evaluation protocols for sensitive AI outputs, focusing on misuse potential, data leakage, and harmful generation risks.
• Author internal policy guidelines for responsible AI development lifecycles, including data provenance tracking, model lineage documentation, and human-in-the-loop review workflows.

• Foundational understanding of AI/ML model types (supervised, unsupervised, LLMs) and their inherent risk vectors including bias amplification, hallucination, and data poisoning.
• Working knowledge of major AI ethics frameworks and governance standards—NIST AI Risk Management Framework, EU AI Act risk tiers, ISO/IEC 42001 AI Management System, and OECD AI Principles.
• Solid grasp of fairness definitions in ML (individual vs. group fairness, conflicting metric trade-offs), privacy concepts (differential privacy, synthetic data), and accountability mechanisms in automated decision-making.

• Python proficiency with libraries such as IBM AIF360, Fairlearn, SHAP, and interpretml for bias auditing and explainability.
• Experience with MLOps or model governance platforms (MLflow, Weights & Biases, or similar) for model lineage and documentation.
• Proficiency in technical documentation and policy writing using Markdown, LaTeX, or Confluence with ability to produce audit-ready deliverables.
• Familiarity with cloud AI services (AWS Bedrock, Azure AI, GCP Vertex AI) and their built-in responsible AI tooling and compliance features.
• Strong research and analysis skills using academic and industry sources (arXiv, Stanford HAI, Partnership on AI, ALBAI guidelines) to stay current on evolving standards.

• A natural instinct to question what a model is *not* telling you—probing silence, missing data, and excluded populations as rigorously as reported metrics.
• Persistent curiosity about the real-world downstream impact of AI decisions, especially in high-stakes domains like healthcare, finance, and cybersecurity.
• Drive to translate abstract ethical principles into concrete, testable engineering checkpoints that development teams can actually implement and verify.

• Competitive annual stipend of ₹12–20 Lakhs with performance-based acceleration.
• Fully remote setup with a home office allowance and monthly Wi-Fi reimbursement.
• Access to dedicated GPU compute credits for experimentation and audit tooling development.
• Direct mentorship from Cehpoint's senior AI engineers, security architects, and governance leads.
• Fast-track conversion to full-time AI Governance Engineer role based on evaluation performance and project impact.