Hey! I'm Srinithi, an AI & Data Science undergrad from Chennai 🇮🇳 who genuinely loves turning messy, real-world problems into clean, working ML systems.
I've spent the past year building things that actually matter — a fraud detection pipeline that cut false negatives by 18%, an agentic research assistant that automates multi-step workflows end-to-end, and a real-time multi-person tracking system using YOLOv8. I don't just train models; I ship them — with Docker, REST APIs, and proper evaluation pipelines.
Right now I'm deep into Agentic AI (LangGraph, LangChain, RAG systems) and I'm fascinated by how LLMs can be wired together to actually do things autonomously. When I'm not coding, I'm organising tech workshops, contributing to open source under GSSoC, or grinding Kaggle competitions.
"Build systems that deliver measurable impact — not just models."
Languages
AI / ML
Data & Analytics
Cloud & DevOps
Databases & Tools
🏢 Zeptoware Systems Pvt. Ltd. — AI & Data Science Intern | Dec 2025 – Jan 2026 | Chennai, India
PythonTF-IDFCNNLSTMDjango REST APIDockerRailway
- Built an end-to-end Fake Job Detection System on 17K+ postings using TF-IDF, CNN, and LSTM — deployed via Django REST API on Railway with Docker, reducing fraudulent listing exposure for real users.
- Reduced false-negative rate by 18% through Bidirectional LSTM and class-weight tuning on a highly imbalanced dataset, directly improving business reliability of the fraud detection pipeline.
- Benchmarked 4 ML models with automated confusion matrix, ROC, and AUC reporting for comprehensive model evaluation.
🌐 Infosys Springboard — AI & Data Science Virtual Intern | Sep 2025 – Nov 2025 | Remote
LangGraphLangChainGemini LLMTavilyScikit-learnPython
- Architected an Agentic Research Assistant using LangGraph state machines, LangChain, Gemini LLM, and Tavily real-time web search — automating multi-step research workflows end-to-end.
- Improved recommendation model accuracy by 20% through Scikit-learn collaborative filtering pipeline optimisation, enhancing personalisation quality for end users.
- Designed modular, reusable agentic components enabling scalable multi-step AI workflows.
✈️ Airports Authority of India (AAI) — Technical Intern | Jul 2025 – Aug 2025 | Chennai, India
PythonSQLData PipelinesReporting
- Streamlined airport data management workflows by building structured data pipelines, improving departmental reporting efficiency.
- Reduced manual processing overhead across operations through automation of data collection and formatting tasks.
- Delivered structured reports improving cross-departmental data visibility and decision-making.
| Project | Stack | Highlights |
|---|---|---|
| 🔍 Fake Job Detection System | Python TF-IDF CNN LSTM Django Docker Railway |
End-to-end fraud detection on 17K+ postings · 18% false-negative reduction · Live production API · Automated ROC/AUC reporting |
| 🤖 Agentic AI Research Assistant | LangGraph LangChain Gemini API Docker Compose LangSmith |
Multi-step agentic pipeline with real-time web search · Full execution tracing via LangSmith · Fully containerised for scalable deployment |
| 👁️ Multi-Person Tracking & Behaviour Analysis | YOLOv8 OpenCV Python |
Real-time multi-subject tracking across video frames · Bounding-box overlays with confidence scoring · Applicable to retail analytics & security monitoring |
| Achievement | Details | |
|---|---|---|
| 🏆 | NPTEL ELITE — Machine Learning | Top 25% Nationally · NPTEL SWAYAM Certification |
| 🌍 | GSSoC 2025 | Open-source contributor to AI/ML repositories · GirlScript Summer of Code |
| 🇮🇳 | Smart India Hackathon 2025 | Shortlisted (college level) · AI & Data Science category |
| 🎤 | College Expo 2025 | Selected from 50+ project teams to present AI-Powered Research Assistant |
| 📊 | Kaggle Expert | Ranked among top contributors · Multiple ML competitions & notebook challenges |
| 🎓 | Microsoft Certification | Fundamentals of Generative AI |
| 🧠 | Google & Kaggle | 5-Day AI Agents Intensive Course |
| 📐 | Coursera | Mathematics for Machine Learning |
| 🤝 | Joint Secretary, IEI | Organised 5+ technical workshops attended by 200+ students |
| Degree | Institution | Year | Score |
|---|---|---|---|
| B.Tech — Artificial Intelligence & Data Science | Jerusalem College of Engineering, Chennai | 2023 – 2027 | CGPA: 8.5 / 10.0 |
Relevant Coursework: Machine Learning · Deep Learning · NLP · Data Structures & Algorithms · Statistics · Database Management · Computer Vision
🧱 Agentic AI → Multi-agent orchestration, tool-use, memory systems
🚀 MLOps → CI/CD for ML, model monitoring, deployment pipelines
📚 RAG Systems → Vector databases, retrieval strategies, hybrid search
🔬 Advanced NLP → Fine-tuning LLMs, prompt engineering, RLHF
☁️ Cloud ML → Scalable training, serving infrastructure
