Available for Research & Engineering Roles · 2026

Yashraj
Sharma

AI/ML Engineer specializing in Computer Vision, Deep Learning & Explainable AI. Published IEEE researcher building systems that achieve up to 97% accuracy. Currently pursuing B.Tech at Institute of Engineering and Management, Kolkata.

Scroll

Builder of
intelligent systems

I'm a final-year Computer Science (IoT, CS & BT) student at the Institute of Engineering & Management, Kolkata< — where theory meets real-world machine learning challenges.

My work sits at the intersection of deep learning and Artificial Intelligence. From predicting gestational diabetes with 97% accuracy using Explainable AI, to verifying faces in adverse weather with Siamese Networks — I build systems that don't just predict, but explain.

I've published at IEEE INDICON 2025 and in a Springer Nature volume. When I'm not training models, I'm serving as Treasurer of the IEEE CIS Society Student Chapter.

97%
Peak model accuracy
2
Research publications
9.12
CGPA (B.Tech)
3+
Research internships
Jun – Aug 2025
Summer Research Intern
IEM Kolkata · Siamese Networks for Face Verification
Apr – Jun 2025
Machine Learning Intern
Feynn Labs · EV Market Segmentation & Crop Disease AI
Dec 2024 – Feb 2025
Winter Research Intern
IEM Kolkata · Gestational Diabetes Prediction + XAI
2025
IEEE INDICON Publication
Siamese-Based Architecture for Face Verification
2026
Springer Nature Publication
ML Framework for Gestational Diabetes with XAI
2023 – Present
B.Tech @ IEM Kolkata
Computer Science · IoT, CS & BT · CGPA 9.12

Technical
Arsenal

🧠 AI / Machine Learning
CNN Computer Vision Neural Networks Transfer Learning Data Science Artificial Intelligence Backend Development
Deep Learning Frameworks
TensorFlow PyTorch Keras Hugging Face Transformer Models NLP Pipeline YOLO OpenCV
🛠 Backend & APIs
Python FastAPI Flask REST APIs SQL Java C
🚀 MLOps & Cloud
Docker GCP Model Deployment Real-time Inference Git NumPy Pandas Scikit-learn Matplotlib

Featured
Work

01 / Computer Vision

Fish Freshness Detection System

Deep learning-based freshness detection using CNN integrated with Grad-CAM for visual interpretability and fuzzy logic for nuanced decision-making. Deployed via Docker with REST API backend.

Impact: Real-time inference with scalable deployment pipeline. Grad-CAM enables visual explainability for quality decisions.
CNN Grad-CAM Fuzzy Logic Docker REST API Python
02 / Computer Vision

Siamese Face Verification Network

Siamese Neural Network for face verification achieving 93% accuracy under adverse environmental conditions (fog, rain, low-light). CNN-based feature extraction with robust data augmentation pipelines.

Impact: Published at IEEE INDICON 2025. Robust to environmental noise and distortions.
Siamese Networks CNN Data Augmentation TensorFlow OpenCV
03 / NLP

Sentiment Analysis Pipeline

End-to-end binary sentiment analysis on movie review datasets. Designed full NLP pipeline including text cleaning, tokenization, vectorization, and transformer-based architectures from Hugging Face.

Impact: Improved contextual representation using BERT-style transformers; validated via precision & recall metrics.
Hugging Face Transformers NLP Python Scikit-learn
04 / Healthcare AI

Gestational Diabetes Prediction

Early-stage GDM prediction using ensemble and meta-learning techniques, achieving 97% predictive accuracy. Integrated Explainable AI for clinical transparency and interpretability.

Impact: Published in Springer Nature 2026. Clinical interpretability via XAI bridges the gap between ML and medical practitioners.
Ensemble Learning Meta-Learning Explainable AI Python Scikit-learn
05 / ML Analytics

Indian EV Market Segmentation

Market segmentation analysis on Indian EV market datasets using clustering and supervised learning. Identified key customer behavior groups to inform business strategy.

Impact: Delivered at Feynn Labs Consultancy. Actionable segments for EV adoption marketing.
K-Means Supervised Learning Pandas Matplotlib Python
06 / Computer Vision

Crop Disease Detection Pipeline

Deep learning-based crop disease classification achieving 95% accuracy. Optimized preprocessing and feature extraction workflows to improve classification consistency for agricultural AI.

Impact: 95% validation accuracy on crop disease classification — supporting precision agriculture applications.
CNN Transfer Learning PyTorch OpenCV Data Augmentation

Publications &
Research

📡 IEEE INDICON 2025 · Conference Paper

Siamese-Based Deep Architecture for Face Verification and Gender Detection in Adverse Weather Condition

Nandi, Apurba et al. (incl. Yashraj Sharma)

A deep learning architecture leveraging Siamese Networks to robustly verify faces and detect gender under challenging adverse weather scenarios including fog, rain, and low-light conditions. The system employs CNN-based feature extraction with data augmentation pipelines to achieve 93% accuracy despite significant environmental noise and image distortions.
📚 Springer Nature Switzerland · Book Chapter 2026

A Comprehensive Machine Learning Framework for Gestational Diabetes Prediction and Risk Analysis Utilizing Explainable AI

Sharma, Yashraj et al. — In: Advanced Computational and Communication Paradigms, Ed. Laia Subirats et al.

Presents an ensemble and meta-learning framework for early-stage gestational diabetes mellitus prediction, achieving 97% predictive accuracy. The work integrates Explainable AI (XAI) techniques to enhance clinical transparency, enabling healthcare practitioners to understand model decisions. Published in Springer Nature Switzerland, pp. 192–202, ISBN: 978-3-032-13555-1.

Work &
Internships

🔬
Summer Research Intern
Institute of Engineering & Management, Kolkata
Developed and optimized a Siamese Neural Network for face verification achieving 93% accuracy under adverse environmental conditions. Implemented CNN-based feature extraction and data augmentation pipelines to improve robustness against noise and distortions. Work resulted in an IEEE INDICON 2025 publication.
Jun – Aug 2025
🤖
Machine Learning Intern
Feynn Labs Consultancy Services
Performed market segmentation on Indian EV dataset using clustering + supervised learning. Built a deep learning crop disease detection pipeline achieving 95% classification accuracy. Optimized preprocessing and feature extraction workflows.
Apr – Jun 2025
🏥
Winter Research Intern
Institute of Engineering & Management, Kolkata
Developed an early-stage Gestational Diabetes Mellitus prediction model using ensemble and meta-learning techniques, achieving 97% predictive accuracy. Integrated Explainable AI to enhance interpretability and clinical transparency — resulting in a Springer Nature 2026 publication.
Dec 2024 – Feb 2025
🌐
IEEE Student Chapter Leader
IEEE OES & CIS Student Chapters
Serving as Chair of the IEEE Oceanic Engineering Society (OES) Student Chapter and Treasurer of the IEEE Computational Intelligence Society (CIS) Student Chapter. Leading technical initiatives and community engagement for 50+ student members.
2023 – Present

Awards &
Recognition

🥇
1st Place — Intra-College Space Quiz
Secured 1st position in the Intra-College Space Quiz Competition (2024).
🎓
13th Rank — B.Tech Program (2024)
Ranked 13th in B.Tech for academic excellence in the department.
🏆
14th Rank — B.Tech Program (2025)
Ranked 14th in B.Tech for academic excellence in the department.
🤖
AI & Data Science Certification
Jadavpur University (2023) — Foundational AI & data science training.
☁️
Google Cybersecurity Certificate
Coursera (2024) — Professional certification in cybersecurity fundamentals.
🐍
Python for Data Science — NPTEL
NPTEL (2024) — National certification for data science with Python.
📡
IEEE INDICON 2025 — Published Author
Peer-reviewed publication at IEEE's India Council International Conference.
📗
Springer Nature — Book Chapter Author
Published in Advanced Computational and Communication Paradigms (2026).

Let's
Connect

I'm actively seeking research collaborations, AI/ML engineering roles, and internship opportunities. If you're working on something interesting in CV, DL, or XAI — let's talk.

✓ Message received! I'll get back to you within 24 hours.