Building secure, compliant AI systems that survive production reality.
I’m a senior engineer and published IEEE researcher working at the intersection of privacy, governance, and enterprise AI security. My focus is turning research-grade ideas into deployable architectures: privacy-preserving analytics, compliance-aware federated learning, secure LLM operations, and scalable observability for cloud environments.
About
I design systems where security, privacy, and compliance are built-in constraints—especially for AI workloads. My work spans differential privacy, federated learning under HIPAA/GDPR constraints, incident detection for LLM systems using minimal audit logs, and lightweight observability architectures for large-scale cloud clusters.
This site is intentionally structured to be “evidence-friendly” for academic, conference, and EB1A documentation.
Research focus
- Privacy engineering (differential privacy, minimal-logging accountability)
- Cross-border federated learning governance (policy-aware collaboration)
- Secure LLM operations (auditability, insider-risk detection, incident response)
- Observability-driven security and cost-aware AIOps
- Responsible AI architectures for enterprise scale
Enterprise leadership
Lead Engineer | Solution Architect
Cornerstone Building Brands (via Chelsoft Solutions)
- Architect and lead enterprise platforms integrating ERP (JDE), pricing intelligence, and automation pipelines.
- Own security-by-design practices: identity, auditability, data minimization, and operational controls.
- Design for reliability: performance, observability, failure isolation, and maintainability.
- Bridge research and production: translating privacy/security concepts into deployable systems.
Architecture principles
- Least exposure: minimize data movement; reduce logs; enforce governance boundaries.
- Operational clarity: observability that supports incident response and accountability.
- Cost-aware scale: measure signal-to-cost; avoid architecture that can’t be operated.
- Proof-ready: decisions documented for audits, compliance, and executive stakeholders.
Selected publications
PrivBuild-AI: An RL-Powered Framework for Differentially Private Data in DevSecOps
Reinforcement learning adjusts privacy controls dynamically to preserve utility while maintaining privacy guarantees in DevSecOps pipelines.
Compliance-Aware Cross-Border Federated Learning for Security Telemetry Under HIPAA/GDPR
A governance-aligned FL framework that enables collaboration while honoring policy constraints and minimizing cross-border exposure.
Who Prompted What? Privacy-Preserving Incident Detection for LLM Systems Using Minimal Audit Logs
Minimal-audit logging strategy for detecting misuse in enterprise LLM systems while reducing privacy risk and operational overhead.
Scalable AIOps: A Framework for Lightweight Observability and Anomaly Detection in Large-Scale Cloud Clusters
Resource-efficient anomaly detection and observability that balances signal quality with cost and operational maintainability.
Speaking
I speak on responsible AI architectures, privacy engineering, federated learning governance, secure LLM operations, and observability-driven security.
- Responsible AI Architectures (enterprise constraints, risk controls, measurable governance)
- Federated learning under HIPAA/GDPR (policy-aware collaboration)
- Secure LLM systems (auditability, minimal logs, insider-risk detection)
- Scalable observability & AIOps (cost-aware anomaly detection)
Events (examples)
- IEEE conferences (AIAHPC, ComManTel, SoutheastCon)
- OWASP / Security community events
- Regional security conferences and industry meetups
Add confirmed dates + proof links as you finalize schedules.
Peer review & professional service
- Peer review for conference/journal submissions (AI, security, privacy)
- Technical Program Committee participation
- Contributor/reviewer in security and standards communities
Review areas
- Post-quantum cryptography key management (cloud & hybrid)
- AI security and LLM governance
- Federated learning, privacy, and compliance engineering
- Cloud security, DevSecOps, observability
Contact
For speaking invitations, collaborations, peer review requests, or enterprise AI architecture engagements:
Email: bhaskar.b.sawant@gmail.com
Location: Canton, Michigan, USA