I’m Sai. I build the engineering around AI that makes it useful: agents that take action, answers with evidence, and systems you can trust.
Find the evidence.
Cloud engineering. Agent workflows. A research lens on AI reliability.
Years across software,
cloud & applied AI
Provisional patent
applications filed
Data Intelligence & AI
University of South Florida
01 SELECTED WORK
Does a prompt-injection detector understand intent, or react to a few familiar words? I study this with controlled word-level interventions and reproducible evaluation.
Reported experiment · NotInject dataset · lower is better
Methodology & results in the repository
A hackathon platform connecting resume analysis, skill matching, and team formation through coordinated services.
Codex Live Lab connects OpenAI’s Responses API with GitHub repository data to support a live product-building workflow.
Enterprise agent workflows, evaluation infrastructure, and secure knowledge retrieval.
02 INTERACTIVE AGENT LAB
Put my experience through its paces.
Every result links back to the evidence.
Runs in your browser. Your text is never sent to a server.
Ask about my work, or paste a role description to find relevant projects and experience.
SOURCE-GROUNDED · INSPECTABLE · LOCALA transparent, local workflow: recognize topics or a fixed vocabulary of technical skills, rank curated portfolio records, validate source links, and assemble evidence excerpts. Role matching shows explicit keyword overlap only; it does not infer proficiency, understand every requirement, or make a hiring recommendation. No LLM, hidden API calls, or fabricated model reasoning.
03 THE PATH HERE
I’ve worked across enterprise software, cloud infrastructure, and applied AI. Today, I’m focused on the space where model behavior meets real-world engineering.
Sole inventor. Filed through USF Technology Transfer. Research interests include distributed inference and language-model reliability.
PROVISIONAL FILINGS · NOT GRANTED PATENTSM.S. Data Intelligence & Artificial Intelligence
University of South Florida · 2024–2026
B.S. Information Technology
Lovely Professional University · 2015–2019
Building retrieval and agent workflows with approved knowledge sources, access controls, structured outputs, and repeatable evaluation.
Developed enterprise backend services and data pipelines. Automated cloud infrastructure and releases with containers, Terraform, and CI/CD.
Built Python inference APIs and data pipelines to move model predictions from experiments into application workflows.
THE TOOLKIT
OpenAI APIs · LangGraph · LangChain
Amazon Bedrock · RAG · MCP
Python · FastAPI · Java · SQL
PostgreSQL · Redis · Event-driven systems
Docker · Kubernetes · Terraform
AWS · CI/CD · OpenTelemetry
04 LET’S BUILD SOMETHING USEFUL
BASED IN TAMPA, FLI’m exploring AI engineering, applied AI, and software engineering opportunities.
saikrishnamanohar@outlook.com