R Sridarsh
builds AI and ML systems.
I'm a computer science undergrad who loves shipping AI and ML systems — not just training them. I've built a radar detection model at DRDO, an agentic RAG pipeline that triages cybercrime threats, and a handful of side projects that all had one thing in common: they ended up as something a real person could actually use.
Experience
Where the work has actually run
Three internships, three very different domains — radar ML at DRDO, quantum + RL research, and C++ tooling. All shipped, none left in a notebook.
AI Engineering Intern
May 2026 — Jun 2026LRDE, DRDO
On-site · Bengaluru, Karnataka
- Co-developed RD-ViT, a hybrid CNN-Vision Transformer for joint detection and localization of small marine targets in sea clutter, using a donor-target transplantation training scheme for gate-agnostic localization.
- Achieved 97.3% detection probability and 87.6% exact range-gate localization (90.2% within ±1 gate) with zero false alarms, at under 0.8M parameters per head — roughly an order of magnitude smaller than comparable detectors.
Research Intern
May 2025 — Jun 2025Chennai Institute of Technology
On-site · Chennai, Tamil Nadu
- Proposed a hybrid Random Forest + MLP framework to mitigate noise in NISQ-era quantum circuits, reaching a test MSE of 4.76e-5 and outperforming standard Zero Noise Extrapolation.
- Implemented a Soft Actor-Critic reinforcement learning agent inside a digital twin simulation to optimize greenhouse climate control, cutting water and energy consumption 20–30%.
Software Development Intern
Nov 2024 — Dec 2024ADE, DRDO
On-site · Bengaluru, Karnataka
- Engineered a C++/MFC video analysis application using OpenCV to extract UAV flight telemetry and payload metadata, serialized into structured XML.
- Developed a specialized C++ desktop search engine to parse XML DOM trees for rapid parameter-based queries against video clips.
Projects
Models that made it out of the notebook
Agentic pipelines, real-time inference behind FastAPI, and the frontends that make their output legible.
Jun 2026 — Jul 2026
- A local Agentic RAG architecture — LangGraph, ChromaDB, Llama 3.2 — that semantically searches and classifies active financial fraud and smishing payloads.
- FastAPI backend with self-correcting agent nodes that generate context-aware mitigation steps and SOC-grade threat assessments in real time.
Jan 2026 — Feb 2026
- A hybrid ML + Agentic AI system that predicts student dropout risk from academic, financial, and behavioral features.
- Random Forest model deployed via FastAPI for real-time inference, with multi-agent orchestration — Explanation and Counselling agents — producing explainable, personalized interventions.
Mar 2025 — Apr 2025
- An automated data-privacy pipeline built on Microsoft Presidio NLP to detect and redact personally identifiable information.
- Interactive Streamlit interface for processing unstructured datasets, built for compliance-first, privacy-first workflows.
Nov 2024 — Dec 2025
- Integrated the OpenAI API with structured prompt engineering to dynamically generate multi-day, personalized travel itineraries.
- React + Firebase web app rendering AI-generated plans, with Google Places and Maps APIs for interactive destination mapping.
Skills
The stack, end to end
From model to API to interface — the tools I reach for at each layer.
Frameworks & Libraries
Languages
Databases
Tools & Platforms
Education
Foundations
Anna University, Chennai Institute of Technology
Bachelor of Engineering in Computer Science and Engineering
Chennai, Tamil Nadu
Aug 2023 — May 2027
Awards & Achievements
Certifications
Contact
Let's build something real.
Available for AI / applied ML internships from 1 September 2026, 2–6 months, on-site or hybrid in Chennai / Bangalore, or remote.
Designed & built by R Sridarsh · 2026