Hi, I'm Mohd Saqib

Vision-Based Queue Analytics System

Developing an automated system using YOLOv8 for person detection and multi-object tracking to estimate queue metrics from live CCTV feeds.

Evaluating system performance on high-density datasets like ShanghaiTech and building an intuitive real-time dashboard using Streamlit/Flask.

Tech Stack: Python, YOLOv8, OpenCV, Streamlit, Deep Learning

EventFlow: Rental Operations Platform

Architecting a scalable monorepo-based operations platform managing 18+ database tables and 12+ type-safe API routers using oRPC and Drizzle ORM.

Implemented a robust RBAC system (Admin, Manager, Staff) using Better Auth and delivered a high-performance UI in an agile Scrum framework.

Tech Stack: React 19, Tailwind v4, Bun, Drizzle ORM

Quiz Master V2 – Multi-User Exam Portal

Developed a full-stack multi-user web app for quiz management using Flask, VueJS, SQLite, Redis, Celery, and Bootstrap.

Tech Stack: Flask, VueJS, SQLite, Redis, Celery, Bootstrap

ML Regression Project – Kaggle Competition

Built and evaluated multiple ML models (Linear Regression, SVM, KNN, MLP, Ensembles) for a Kaggle regression competition.

Tech Stack: Python, Scikit-learn, XGBoost, Pandas, Matplotlib, Seaborn

Business Optimization for Electronics Retailer

Digitized and analyzed 30+ days of sales & inventory data, enabling structured analysis and proposing actionable improvements.

Tech Stack: Python, Pandas, Excel, Google Sheets, Matplotlib, Seaborn

LLM-based Automation Agent

Built a robust automation agent using GPT-4o-mini that accepts plain-English tasks and executes them via API.

Tech Stack: Python, GPT-4o-mini, Docker, Shell Tools, API Integration

TDS Solver – LLM API for Assignment Questions

Developed a production-ready REST API that answers assignment questions using LLM-based inference.

Tech Stack: Python, GPT-4o-mini, REST API, Vercel, JSON

Canteen Management System

Designed a scalable system to manage real-time orders and inventory across 4 parallel canteens.

Tech Stack: Node.js, Express.js, MySQL, HTML, CSS, JavaScript, Bootstrap

Smart Home Services Platform

Developed a multi-role web application to connect customers with home service professionals.

Tech Stack: Flask, SQLite, Flask-Login, Jinja2, Bootstrap, JavaScript, Chart.js

Visual Question Answering (VQA) System – Image & Video

Built a unified deep learning system that answers natural-language questions from both images and videos using a frame-based approach. Implemented multimodal fusion with BLIP (ViT + BERT) to combine visual and textual features and generate accurate answers.

Designed image and video pipelines where keyframes are extracted from videos, processed independently, and aggregated using confidence-weighted selection. Deployed an interactive web interface for real-time experimentation.

Tech Stack: Python, PyTorch, HuggingFace (BLIP), OpenCV, Flask, HTML, CSS, JavaScript

Academic: Minor Project (B.Tech, Computer Engineering) — Jamia Millia Islamia, 2025–26

Smart Hospital Management System (HMS)

Designed and developed a full-stack, role-based Hospital Management System to streamline doctor, patient, and appointment workflows. The system replaces manual processes with a secure, web-based platform for efficient coordination and record-keeping.

Implemented separate dashboards for Admin, Doctor, and Patient with features such as appointment booking, treatment history, medicine inventory, prescriptions, PDF invoices, reviews, and analytics using charts.

Tech Stack: Flask, SQLite, Jinja2, Bootstrap 5, Chart.js, Flask-Login

Academic: Modern Application Development-I (MAD-I), IIT Madras BS — September 2025