- Built C# and PowerShell automation workflows migrating 100,000+ records
- Reduced manual data transfer by 70%
- Developed internal SQL/.NET reporting tools reducing validation errors by 25%
- Evaluated enterprise AI readiness and contributed to foundational AI policy
anubhav@portfolio — zsh
anubhav@portfolio:~$
█
const anubhav = {
role: "AI Researcher & Applications Developer",
education: "University of South Florida — 4.0 GPA",
focus: ["LLMs", "Enterprise Systems", "Optimization"],
status: "Building systems that solve measurable problems"
};
scroll ↓
about.md
markdown
# About Me Hi, I'm Anubhav Gupta — a Computer Science student at the University of South Florida with a 4.0 GPA, undergraduate AI researcher, and applications developer working at the enterprise level. I operate at the intersection of Artificial Intelligence, systems engineering, and real-world production software. My work spans enterprise automation, large language model research, algorithm design, and scalable web platforms. // I don't just build projects — I build systems that solve measurable problems.
// experience.log
- Persona-Conditioned Behavioral Deviations and Targeted Unlearning in LLMs
- Bias and behavioral deviation across demographic personas
- Power-compliance patterns, interpretability, and mitigation strategies
- Building responsible AI systems that are robust, fair, and aligned
- Solved NP-hard Vehicle Routing Problem using Google OR-Tools
- Fallback routing logic reducing API dependency by 95%
- Integrated OpenStreetMaps, HERE, and Google Maps APIs
- GeoJSON dashboards for real-time analytics
- Managed global research datasets
- Built AI classification pipeline using third-party LLMs
- Reduced manual entry time by 95% through semantic classification
- TA — CDA 3201 (Logic Design) — USF Bellini College of AI, Cybersecurity & Computing
- Tech Chair — Google Developer Group @ USF — Led team of 8 technical leads
- Software Lead Shadow — ACM — DiagnoAI (CNN-based chest X-ray anomaly detection)
// selected_projects
🏆 HackaBull 2025 Track Winner
CarbonCTRL
AI-driven carbon emissions dashboard with Gemini API integration. Improved insight generation by 40%, reduced backend latency by 35%.
ReactTypeScriptSupabaseGemini API
PalmVerse
Real-time sign language translation with 87% model accuracy. MediaPipe + OpenCV + Flask. Reduced processing latency by 15%.
PythonMediaPipeOpenCVFlask
// tech_stack.config
Languages
C++ · Java · Python · C# · SQL · JavaScript
Frameworks
.NET · React · Node.js · Express · MongoDB · Supabase · Angular
AI & ML
LLMs · TensorFlow · CNNs · OR-Tools
Infrastructure
AWS · Linux CLI · Git · Jira · Postman
mission.js
const mission = { build: [ "AI systems that interact with power structures responsibly", "Scalable infrastructure for enterprise systems", "Consumer AI products with real-world utility", "High-performance optimization platforms" ], longTerm: "Intersection of AI, business systems, and large-scale impact" };
connect