01

Future forward
creative
engineer.

blending machine learning expertise
with engineering clarity to build
systems that stand out, move fast, and
stay unforgettable.

Based in Jaipur, building globally.
I Deconstruct, Rebuild, and Engineer Your Ideas

About Me

I'm a B.Tech Computer Science student at Government Engineering College, Jaipur (CGPA: 8.90) with a strong interest in Machine Learning, Artificial Intelligence, and Software Development.

I enjoy building end-to-end AI applications that combine machine learning models, backend systems, and user-facing products. My experience includes developing fraud detection systems using LightGBM and XGBoost, creating AI-powered meeting intelligence platforms with speech recognition and LLM integrations, and building cross-platform mobile applications using Flutter.

Currently, I'm focused on Deep Learning, Computer Vision, NLP, and building real-world AI products that solve practical problems.

Mission

To engineer intelligent systems that solve high-stakes real-world problems — making machine learning not just accurate, but reliable, scalable, and deployable in production environments.

Vision

To become a leading ML/AI engineer who bridges the gap between research and production, building systems that create measurable impact at scale across fraud prevention, computer vision, and language intelligence.

Work Experience

Full-time Internship June 2026 – Present

Machine Learning Intern

Sun Shine Automation India Limited (SSAIL)

📍 Jaipur, Rajasthan, India · On-site

Key Responsibilities
  • Developed a real-time PPE detection system using YOLOv8, identifying safety helmets and gloves at 30+ FPS via live webcam feeds.
  • Solved model degradation (catastrophic forgetting) by implementing an ensemble architecture that runs multiple neural networks simultaneously to merge predictions.
  • Programmed custom object tracking using Intersection over Union (IoU) and temporal smoothing to filter webcam noise and stabilise bounding boxes.
  • Managed the full data pipeline — converting, merging, and normalising XML and YOLO datasets for PyTorch model training on an RTX 4050 GPU.
Key Achievements
  • Achieved 30+ FPS real-time inference on live webcam feeds, meeting production deployment requirements for on-site industrial safety monitoring.
  • Resolved catastrophic forgetting via ensemble fusion, significantly improving multi-class detection stability across varying lighting and occlusion conditions.
Internship June 2025 – July 2025

Android App Development Intern

Centre for Electronic Governance (CEG)

📍 Jaipur, Rajasthan

Key Responsibilities
  • Developed responsive cross-platform Android applications using Flutter and Dart with adaptive UI components for multiple screen sizes.
  • Designed and integrated end-to-end app workflows ensuring seamless user experience across all device form factors.
  • Collaborated on backend integrations, state management, and performance optimisation for real-time application features.
Key Achievements
  • Delivered production-ready Flutter applications deployed within a government digital services initiative serving citizens across Rajasthan.
  • Optimised real-time feature performance, reducing UI lag and improving app responsiveness across low-end Android devices.
Industrial Training July 2024 (1 month)

Industrial Trainee — Machine Learning & IoT

MIIC, MNIT Jaipur

📍 Jaipur, Rajasthan

Key Responsibilities
  • Completed industry-focused training in ML and IoT with hands-on exposure to predictive analytics on real industrial datasets.
  • Worked on data preprocessing, feature engineering, and deployment-oriented ML pipelines for industrial use cases.
Key Achievements
  • Applied ML techniques to real sensor data, producing deployment-ready predictive models for industrial monitoring scenarios.
  • Gained practical exposure to end-to-end ML workflows in an industrial research environment at one of India's premier NIT institutions.

My Projects

Computer Vision / MLOps / Full-Stack AI

EdgeSafety-AI — Real-Time PPE & Safety Detection System

A production-oriented PPE compliance monitoring system built on a 3-model YOLOv8 ensemble with a 14-class unified detection registry. The system identifies helmets, gloves, vests, goggles, masks, and hazardous events in real time at 30+ FPS via live webcam feeds, with a FastAPI web dashboard for live streaming and static image analysis.

  • Engineered a 3-model ensemble with custom cross-model IoU/IoM NMS filter to resolve overlapping detections across all 14 classes.
  • Achieved 94.3% mAP@50 (helmet) and 94.1% mAP@50 (glove) after 100-epoch AdamW training on RTX 4050.
  • Implemented 30-frame temporal smoothing buffer with IoU-based tracking to eliminate label flickering.
  • Built complete FastAPI web dashboard with real-time MJPEG streaming, dynamic PPE target isolation, live detection stats, and interactive training metric charts.
PythonYOLOv8PyTorchFastAPIOpenCVCUDAChart.js
Machine Learning / Fraud Detection

ML Mule Account Detection System

A high-accuracy fraud detection pipeline built on large-scale financial datasets, targeting mule account classification. The system combines advanced feature engineering with an ensemble of gradient boosting models to achieve near-production AUC scores.

  • Engineered 180+ behavioural and transactional features from raw financial data using custom preprocessing pipelines.
  • Developed a LightGBM and XGBoost ensemble achieving 0.983 AUC — near-perfect for binary fraud classification.
  • Built a modular pipeline supporting feature selection, model training, and threshold-optimised inference.
  • Recognised in the National Fraud Prevention Challenge (NFPC) for the approach.
PythonLightGBMXGBoostScikit-learnPandasFastAPI
AI / NLP / Backend

Samwaad AI — Meeting Intelligence Platform

A real-time AI platform that transforms raw meeting audio into structured intelligence — transcription, summarisation, sentiment analysis, and action-item extraction — with automated post-meeting workflows integrated into Google Calendar and email.

  • Integrated Google Speech-to-Text and Gemini AI for accurate speaker diarization and contextual meeting summarisation.
  • Automated post-meeting workflows via Google Calendar and SendGrid, eliminating manual follow-up effort.
  • Built scalable backend APIs and real-time processing workflows for concurrent meeting intelligence sessions.
  • Designed a system for extracting structured action items and sentiment signals from unstructured transcripts.
PythonGemini AIGoogle STTFastAPISendGridNLP
Mobile / AI / Data Visualisation

AI-Driven Finance Management App

A cross-platform personal finance application that combines expense tracking and spending visualisation with AI-driven financial insights, built for both Android and iOS using Flutter.

  • Developed cross-platform mobile app with responsive dashboards for real-time expense tracking and category-level analysis.
  • Integrated AI-driven insight engine to surface personalised spending patterns and financial recommendations.
  • Designed analytical visualisations that translate raw transaction data into clear, actionable financial summaries.
FlutterDartFirebasePythonAI/MLData Viz

Other Projects & Archives

More repositories, experiments, and earlier work.

View Archive →

My Education

B.Tech in Computer Science & Engineering

CGPA: 8.90

Government Engineering College, Jaipur

August 2023 – May 2027 · Full-time

  • Specialisation in Machine Learning, Artificial Intelligence, and Data Structures & Algorithms.
  • Consistent high academic performance (CGPA 8.90) alongside active project development and hackathon participation.
  • Built a portfolio of production-oriented ML and mobile development projects while completing 8 industry certifications in parallel.

My Certifications

NVIDIA Certificate
NVIDIA

Applications of AI for Anomaly Detection

Stanford Certificate
Stanford & DeepLearning.AI

Machine Learning Specialization

Google Certificate
Google & Kaggle

5-Day AI Agents Intensive Course

LinkedIn Learning CV Certificate
LinkedIn Learning

Building Computer Vision Apps (Python)

LinkedIn Learning Data Viz Certificate
LinkedIn Learning

Python for Data Visualization

Anthropic Certificate
Anthropic

AI Fluency: Framework and Foundations

Udemy Dart Certificate
Udemy

The Complete Flutter Development Bootcamp with Dart

Achievements & Recognition

Competing and winning at national level — from fraud prevention challenges to AI hackathons at India's top engineering institutions.

0+ Years Experience
0+ Certifications
0 Projects
0 Awards
2025
Professional

5th Rank — Sphinx'25 National Hackathon

MNIT Jaipur

Secured 5th position at Sphinx'25, a national-level hackathon hosted at MNIT Jaipur, competing against teams from across India with an innovative AI-driven solution.

2025
Professional

Certificate of Appreciation — MUJ HackX 3.0

Manipal University Jaipur

Awarded for outstanding technical innovation and rapid prototyping execution at MUJ HackX 3.0, recognised for delivering a working AI prototype under hackathon time constraints.

2025
Professional

National Fraud Prevention Challenge (NFPC) — Recognition

National Fraud Prevention Challenge

Recognised in the NFPC for developing an AI-powered fraud detection solution that demonstrated high-accuracy mule account classification on real financial datasets (0.983 AUC).

Let's connect.

Send a message
My Resume