Krishnendu Dasgupta

Applied AI security and technical AI safety researcher-engineer with over 15 years of experience across machine learning, generative AI, healthcare AI, and cybersecurity-adjacent systems. Core research focuses on evaluating and hardening open-weight AI systems that operate over sensitive data through enforceable runtime controls — privacy-preserving local inference, prompt-injection resistance, policy-as-code governance, deterministic orchestration, least-privilege tool access, and human-in-the-loop oversight. Contributor to five NIST GenAI evaluation tracks (code generation, text discrimination, text generation, image discrimination, image generation), NIST ARIA 0.1 risk-and-impact assessment, and NIST AI Agent Security standards development (Docket NIST-2025-0035).

Selected as a United Nations Development Programme vetted consultant for AI tracks under the ExpRes Roster for GPN deployments (2025). Trained and invited speaker at OWASP Italy Day, KubeCon + CloudNativeCon Europe, Linux Foundation AI Summit, and international AI safety and data science conferences.

Founder & Independent AI Security Researcher - AXONVERTEX AI - Bengaluru, India

(2023-05)

• Lead independent research on AI security, technical AI safety, open-weight model evaluation, private compute, and decentralized inference for high-risk domains including healthcare and cybersecurity.
• Selected as United Nations Development Programme vetted consultant for Artificial Intelligence tracks under the ExpRes Roster for GPN deployments (2025).
• Participated in five NIST GenAI evaluation tracks (code, text generation, text discrimination, image generation, image discrimination) and NIST ARIA 0.1 risk-and-impact evaluation, producing system descriptions, evaluation pipelines, and local inference infrastructure.
• Submitted detailed response to NIST/CAISI AI Agent Security RFI (Docket NIST-2025-0035) with agent-specific threat modeling, deployment controls, monitoring, rollback, identity, and standards recommendations.
• Develop secure AI agent architectures with policy-as-code, PII/secret redaction, deterministic orchestration, sandboxed tool use, audit trails, and human approval workflows.
• Build healthcare AI systems across FHIR-grounded graph reasoning, TrialBridge clinical trial intelligence, RxNorm/Gemma normalization, private medical relay infrastructure, and multimodal clinical agents.
• Created and delivered OWASP Italy Day 2026 training on building, securing, and deploying AI agent swarms in trustless decentralized ecosystems (Cagliari, Italy).

Head of Machine Learning → Data Consultant - Mondosano GmbH - Remote / Germany

(2022-07 - 2024-10)

• Led development of graph machine learning platform integrating patients, diseases, drugs, symptoms, and clinical trials for clinical trial recruitment and matching.
• Built AI-powered clinical trial recommendation systems and generative AI workflows for disease ontologies and clinical research use cases.

Technical Head – AI / Ex-Co-Founder - DoConvid AI - India

(2020-06 - 2022-05)

• Built privacy-preserving decentralized ML ecosystem for asthma diagnosis/prognosis and medical imaging platform for lung abnormality detection in X-rays.
• Led NVIDIA Inception / grant-supported work using PyTorch, MONAI, and GPU-accelerated clinical imaging infrastructure.

Lead Senior Consultant – AI & ML - Thoucentric - India

(2018-10 - 2020-06)

• Built machine learning platforms for demand projection and demand sensing; led ML deployment teams for Fortune 500 supply-chain clients.

Senior Consultant - PricewaterhouseCoopers - India / Africa

(2017-01 - 2018-10)

• Built AI implementation platforms across healthcare, banking, and cybersecurity; led AI initiative in Kenya with Microsoft Africa and PwC Africa.
• PwC Innovation Leap Award winner — Best Innovation and Use of AI & ML. Represented PwC India in global internal AI hackathon-winning team.

Principal Consultant / Multiple Roles - NTT DATA - India / Global Delivery

(2010-07 - 2017-01)

• Built global supply-chain analytics, statistical modeling, and ML systems for major FMCG clients across 40+ countries.
• Delivered Project Sunset (multi-level escalation system) and Early Warning Write-Off system saving more than INR 15M annually.

Bachelor of Engineering - Computer Science - Visvesvaraya Technological University (2005 - 2009)

Technical AI Safety Course - BlueDot Impact (2026 - 2026)

Stanford Code in Place — Section Leader (CS106A) - Stanford University / Stanford Engineering, Center for Global & Online Education (2024 - 2026)

Professional Program - Artificial Intelligence and Graph Machine Learning - Stanford University (2024 - 2024)

Certificate - Medical Statistics - Stanford University (2023 - 2024)

Artificial Intelligence in Healthcare - MIT Executive Education (2020 - 2020)

Making Data-Driven Decisions - MIT Professional Education (2020 - 2020)

Negotiations - MIT Professional Education (2019 - 2019)

Innovation and Entrepreneurship Bootcamp - Massachusetts Institute of Technology (2018 - 2018)