ANKIT GHOSH
Deep Learning · Agentic Systems · CV
AI/ML Engineer · B.Tech AI · SRM IST 2027
Deep Learning · Agentic Systems · CV
AUTHENTICATED CREDENTIALS
Specializing in curriculum LLM training, computer vision pipelines, & high-throughput backend architecture.
Architecting production AI systems — from LLM curriculum training
to distributed inference orchestration at scale.
About Me
I'm an AI/ML Software Engineer specializing in Deep Learning, Computer Vision, and Agentic AI systems. My work spans curriculum-based LLM training, custom object detection, distributed inference orchestration, and production-grade ML pipelines — all containerized and battle-tested.
Currently in my 3rd year at SRM Institute of Science & Technology, Kattankulathur, pursuing B.Tech in Artificial Intelligence (CGPA 8.20/10). I build systems the way they should be built — scalable, observable, and production-shaped from day one.
Projects
Mobile Payment & Digital Wallet Platform
A secure, fast, and lightweight digital wallet & mobile payment application designed for instant peer-to-peer transfers, transaction tracking, and seamless payment management — working fully offline without an internet connection.
Smart Pantry & Grocery Inventory
Photograph a receipt or shelf — item name, category, and estimated expiry extracted by a CNN + OCR pipeline. The resulting inventory is queryable by any MCP-compatible AI client. Answers: what's in the house, and what's about to go bad.
Research
01
PublicationZenodo · Nov 2025
Developed an adaptive, time-series portfolio-optimization framework using Robbins–Monro stochastic approximation, enabling real-time weight updates under non-stationary market conditions with EWMA-based volatility forecasting. Achieved 22.8% annualized return and Sharpe ratio ≈ 1.0 over 10+ years of financial data.
Read Paper ↗02
Research ProjectComputer Vision · Autonomous Perception
Monocular depth-based vehicle detection using geometric deep learning. Implements neural architecture for autonomous perception without stereo cameras, exploring depth estimation and object localization from a single RGB feed.
View Repo ↗03
AI ResearchIndependent Research · Financial AI Evaluation · 2026
Empirical alignment study comparing a Groq 70B Teacher Judge and a sub-millisecond local heuristic evaluator against a 3-annotator human consensus ceiling across 200 financial probe completions. Explores RLAIF viability via Fleiss's Kappa, Quadratic Cohen's Kappa, and Pearson/Spearman correlation — targeting BNY Mellon, UBS, Goldman Sachs AI teams.
View Repo ↗04
Agentic AI ResearchIndependent Research · Jan 2026 – Present
An original agentic training framework where a 70B critic model autonomously generates training data targeting weaknesses found in the fine-tuned model's outputs — turning evaluation directly into the next round's dataset in a fully closed loop.
View Repo ↗Experience
Intern — Software Developer (Full-Time)
📍 Sector V, Bidhannagar, Kolkata — Bengal Intelligent Park
Education
B.Tech — Artificial Intelligence
2023 – 2027
CGPA: 8.20 / 10
12th — Computer Science, Alwar, Rajasthan
2022 – 2023
Score: 7.78
10th — Computer Science, Janai, West Bengal
2020 – 2021
Score: 8.89
Achievements
NitroStack · Aug 2026
Built a winning multi-agent MCP server automating industrial recovery — real-time anomaly detection, predictive maintenance, and ERP/MES/IoT integration.
View Certificate ↗Contact
Open to internships, research collaborations, and interesting AI/ML problems. Feel free to reach out — I usually respond within 24 hours.