Open to Opportunities — United States

Sadhana Sainarayanan

Cloud platform Security Analyst working at the intersection of SAP BTP, agentic AI, and cloud security automation. Currently working on personal projects to build tools that make AI systems observable and accountable.

View Projects Get in Touch

Background

Building at the edge of
cloud security and AI systems

I'm an Analyst at TCS, embedded at a large tech client, where I work in SAP BTP Cloud Platform Foundations and Security, Python Automation, and Claude SDK based agent development. My dual master's from Carnegie Mellon University grounds my work in systems thinking. I approach cloud security the way an engineer approaches infrastructure: with rigor, auditability, and a bias toward automation over process.

My personal projects include building a tool stack for agentic AI accountability and security that detect behavioral divergence, sign execution logs, and surface when AI agents act differently when they think no one is watching.

DEF CON

Cloud Village speaker, August 2026

Agentic AI tools in active development

Capabilities

What I work with

Cloud & Platforms

AWS Azure GCP SAP BTP SAP Data Intelligence Docker Kubernetes

AI / ML

LLM APIs RAG AI Agents Multi-agent Systems MCP LoRA NLP (RNN / Transformers) Ensemble Methods Gurobi Optimization

MLOps & Observability

MLflow OpenTelemetry Dynatrace Splunk Grafana Kibana FastAPI CI/CD (Jenkins, GitHub Actions)

Languages

Python R SQL Node.js

Tools

Tableau Power BI Spark Postman Jira Confluence

Certifications

GCP Professional ML Engineer (Exp Mar 2026) Azure AI Engineer Associate (Exp Oct 2026)

Work

Projects

Active Build · DEF CON 34 Cloud Village · Aug 7 2026 · Las Vegas

Trust Fall: How Agentic AI Inherits Your Cloud's Worst IAM Habits

Research examining how AI Agents inherit and amplify least-privilege violations already present in cloud IAM configurations. Co-presented with Aravind Pallavoor.

AWS IAM Agentic AI MCP Security Research

Active Build · AISF Vegas · Aug 6 2026 · Las Vegas

AI Agent Specification Gaming: How Agents Weaponize Compliance

Three-layer threat taxonomy for specification gaming where AI agents satisfy the letter of their instructions while violating intent. Includes a spec auditor tool and resolution tooling with live demos using an AWS Emulator. Co-presented with Aravind Pallavoor.

AI Control Specification Auditing Agentic Security Python

github.com/SadhanaSai/spec-gaming-agents

Active Build

AWS Security Finding Analyst Agent

LangGraph + Claude API agent that triages GuardDuty and Security Hub findings, enriches context, and surfaces prioritized remediation paths. Built to reduce analyst toil on high-volume finding queues.

LangGraph boto3 Claude API GuardDuty

Active Build

BehaviorProbe / researchgapfinder

BehaviorProbe for model drift detection across versions

Python NLP LLM Accountability

github.com/SadhanaSai/behaviorprobe

Career

Experience

Nov 2023 — Present TCS · FAANG Client

Analyst — ML & Cloud Platform Engineering

Built an AI agent with MCP-enabled multi-client tooling to automate security onboarding, configuration updates, and incident resolution across 50+ cloud projects, reducing manual effort for 60+ developers
Improved incident identification efficiency by 50% by implementing auto-instrumentation with Dynatrace and OpenTelemetry and establishing proactive threshold-based alerting for production Python applications
Accelerated security configuration deployment by 80% (5 min to under 1 min) by defining DevOps automation requirements and troubleshooting pipeline code across 50+ projects
Drove security configuration setup across 50+ projects

Dec 2021 — Nov 2023 TCS · FAANG Client

Analyst — Machine Learning Operations (MLOps)

Reduced data ingestion time by 80% for 100M+ row production datasets by designing a custom pipeline architecture that worked around SAP Data Intelligence platform limitations, improving model refresh SLAs
Automated ML lifecycle — monitoring and retraining — for a time-series forecasting model serving a retail client, reducing manual intervention
Built and deployed NLP classification pipelines and LLM POCs; integrated event-driven microservices with ML pipelines
Delivered retail optimization models using ensemble methods and Gurobi

Oct 2021 — Jan 2022 City of Pittsburgh

Mobility Data Analytics Intern — Dept. of Mobility & Infrastructure

Data analysis for Public Transportation and mobility equity

Jul 2021 — Oct 2021 WSP USA

Systems Analysis Group Intern — Transportation Modeling and Data Science

Transportation modeling and data science

2020 Carnegie Mellon University

Research Assistant — Civil & Environmental Engineering

Micromobility data analysis under Prof. Corey Harper — analyzed potential for micromobility to replace short urban car trips and reduce emissions using Seattle household trip survey data

Education

Academic background

Carnegie Mellon University

M.S. Civil & Environmental Engineering
Advanced Infrastructure Systems Concentration

Pittsburgh, PA · 2021

GPA 3.73 (Dual Degree Consolidated)

Carnegie Mellon University

M.S. Engineering & Technology Innovation Management
Analytics Concentration

Pittsburgh, PA · 2021

SSN College of Engineering

B.E. Civil Engineering

Chennai, India · 2019