N
a
m
a
s
t
e
(
)
;
I
'
m
S
r
i
v
a
t
s
a
K
a
m
b
a
l
l
a
CS graduate from UIC building reliable systems across backend engineering, AI infrastructure, cloud platforms, and automation.
Recent work includes UIC x CCC Intelligent Solutions MetARAG, UIC x TransUnion ECI Pipeline, FuzzBench research infrastructure, and Kubernetes DevSecOps platforms.
RAG-based document intelligence for querying large policy and agreement collections with source-grounded answers.
AI-powered ecosystem change intelligence for Android security, CVE, fraud, and risk-analysis workflows.
Reproducible automated software testing infrastructure on AWS with Docker, Bayesian estimators, and benchmark pipelines.
Three-tier Kubernetes platform with Terraform, Jenkins, ArgoCD, security gates, Prometheus, and Grafana.
Co-invented a patented AI-enabled library automation system using cameras, RFID, and access-control workflows.
Hello, I'm Srivatsa Kamballa - a Computer Science graduate from UIC focused on software engineering across AI systems, backend infrastructure, cloud platforms, and automation.
My work includes RAG-based document intelligence with CCC Intelligent Solutions, ecosystem change intelligence with TransUnion, research infrastructure on AWS, and Python/SQL data automation.
I care about building systems that are practical, maintainable, observable, and useful beyond the first prototype.
Problem: Make large policy and agreement PDFs searchable through natural-language Q&A with source context.
Built: Parsing, chunking, metadata enrichment, embeddings, retrieval tuning, and evidence-backed answer workflows.
Stack: Python, LangChain, OpenAI APIs, Docker, Kubernetes, PostgreSQL, pgvector.
Problem: Track Android ecosystem changes across security, CVE, policy, and developer sources for risk workflows.
Built: DeltaRAG and Graph-RAG pipeline that turns cross-source evidence into structured risk insights and action tickets.
Stack: Python, FastAPI, PostgreSQL, pgvector, NetworkX, Groq LLM, AWS Lambda, Next.js, Vercel.
Problem: Rebuild a fragile fuzzing research environment into a reproducible benchmark pipeline.
Built: Dockerized FuzzBench workflow with a custom LLVM toolchain, standardized experiment scripts, and Bayesian estimators.
Stack: Google FuzzBench, Docker, LLVM, AWS, Python, Bash.
Problem: Understand production-style cloud deployments across infrastructure, CI/CD, GitOps, security, and observability.
Built: AWS EKS platform with infrastructure-as-code, automated releases, security scanning, rollback workflows, and monitoring.
Stack: AWS EKS, Terraform, Jenkins, ArgoCD, Trivy, SonarQube, Prometheus, Grafana.
Mar 2025 - May 2026
Chicago, Illinois
Sep 2024 - Mar 2025
Chicago, Illinois
May 2023 - Jun 2024
Bangalore, India
Aug 2022 - May 2023
India
Built a production-style three-tier application platform on AWS EKS to understand how modern deployments work end to end.
Provisioned infrastructure with Terraform, implemented GitOps workflows with ArgoCD, automated CI/CD through Jenkins, and integrated Trivy, OWASP scanning, SonarQube, Prometheus, and Grafana across the release workflow.
Rebuilt automated software testing research infrastructure using Google's FuzzBench, Docker, AWS, Python, Bash, and a custom LLVM toolchain.
Implemented Bayesian estimators for predicting path discovery in fuzzing, evaluated results across real-world benchmarks, and fixed a scale mismatch between single-thread predictions and multi-fork ground truth.
Worked on a UIC industry collaboration with CCC Intelligent Solutions to build a RAG-based document intelligence platform for querying large document collections in natural language.
Built the data and retrieval pipeline for 40+ PDFs, including parsing, chunking, metadata enrichment, embeddings, retrieval tuning, and source-grounded response generation using Python, LangChain, OpenAI APIs, Docker, Kubernetes, PostgreSQL, and pgvector.
Worked on a UIC industry collaboration with TransUnion to build an AI-powered ecosystem change intelligence pipeline for monitoring Android security, CVE, and risk-related signals.
Designed a DeltaRAG and Graph-RAG workflow using Python, FastAPI, PostgreSQL, pgvector, NetworkX, Groq LLM, AWS Lambda, Next.js, and Vercel to turn cross-source evidence into structured risk insights and action tickets.
Built an AI-powered movie recommendation system that personalizes film suggestions using user interaction data.
Implemented collaborative filtering, content-based filtering, data cleaning, feature engineering, API integration, and MLOps-style practices including containerization, CI/CD, monitoring, and model-serving workflows.
Built an end-to-end encrypted messaging application inspired by Telegram's MTProto 2.0 protocol for an advanced security course.
Implemented client-server messaging, secure session handling, Diffie-Hellman key exchange, AES-based encryption, message authentication, and forward secrecy.
Team project where we built a multi-agent AI system that catches fraud in real time — it looks at transaction patterns, user behavior, and risk signals together rather than in isolation.
My part was designing the AI agent architecture and the fraud analysis pipeline that connects detection across multiple signals. The goal was catching things rule-based systems typically miss.
Designed and implemented a domain-specific language for fuzzy logic operations in Scala.
Built the parser, interpreter, and optimizer from scratch, with support for fuzzy set operations, logic gates, partial evaluation, and dynamic binding.
Research and engineering work currently in progress.
Building a reproducible fuzzing research pipeline that estimates residual software risk using Bayesian statistical estimators, Google FuzzBench, Docker, LLVM, and large-scale benchmark data.
View GitHubBuilding an ecosystem change intelligence system that monitors Android security, CVE, and policy sources, then turns cross-source evidence into structured risk insights for fraud and risk workflows.
View GitHubBuilding a production-style AWS EKS platform with Terraform, Jenkins, ArgoCD, Trivy, SonarQube, Prometheus, and Grafana to practice secure, observable, zero-manual deployment workflows.
View GitHubDesigned around source-grounded retrieval, metadata enrichment, and response evaluation so answers stay tied to evidence.
Combines graph context, vector retrieval, and LLM agents to convert fragmented security signals into usable risk summaries.
Uses infrastructure-as-code, CI/CD, security gates, deployment automation, and monitoring to model production release workflows.
Python
JavaScript
C++
Java
YAML
Shell Scripting / Bash
React
Node.js
SQL / PostgreSQL
Amazon Web Services
Microsoft Azure
Google Cloud Platform
Linux
Jenkins
ArgoCD
Trivy
Terraform
Ansible
HashiCorp Packer
Docker
Kubernetes
Helm
Istio Service Mesh
Prometheus
Grafana
Tableau
Azure DevOps
GitHub
Jira
Slack
Microsoft Teams
Visual Studio Code
Aug 2024 - May 2026
Chicago, Illinois
GPA: 3.88/4.00
Relevant Coursework:
2020 - 2024
Bangalore, India
GPA: 3.91/4.0
Aug 2025 - May 2026 | Chicago, Illinois
Patent No. 202341071153
Issued: November 24, 2023 | Intellectual Property India
The invention leverages AI-enabled cameras, RFID systems, and automated access control to streamline book issuance and return processes in libraries through intelligent user identification and tracking.
Currently open to full-time roles starting June 2026
Based in Chicago and open to relocating across the United States.
Start a conversationRecommendations from UIC faculty, CCC Intelligent Solutions collaborators, and TransUnion project stakeholders will be added here after they are published on LinkedIn.
Get in touch
I am open to full-time software engineering, AI infrastructure, platform engineering, cloud engineering, DevOps/SRE, and data infrastructure roles starting June 2026.
