AI
M.Tech AI & ML · VIT Vellore

Ahalya
Rajan

Graduate researcher and AI/ML engineer building end-to-end intelligent systems — from trustworthy deep learning to real-time deployed applications. 4× published in IEEE & Springer.

View Projects Download CV
About Me

Building AI that
you can trust

I'm a graduate researcher at VIT Vellore specialising in trustworthy AI, NLP, and deep learning systems. My work spans research prototypes to production-grade deployments — always with an emphasis on reliability, explainability, and performance.

With a strong foundation in data engineering, MLOps, and intelligent automation, I focus on creating AI systems that are not just accurate, but transparent and deployable in real-world environments.

4×
IEEE & Springer Publications
5+
End-to-End AI Projects
2
Leadership Roles
M.Tech
CSE (AI & ML), VIT Vellore
Education
2025 – 2027
M.Tech — Computer Science & Engineering (AI & ML)
Vellore Institute of Technology, Vellore
2021 – 2025
B.Tech — Artificial Intelligence & Machine Learning
Kongu Engineering College, Erode
2021
HSC
Palaniappa Matriculation Higher Secondary School
2019
SSLC
AVP Matriculation Higher Secondary School
Expertise

Technical Skills

Programming
PythonSQL JavaC
Machine Learning & Deep Learning
MLDeep Learning NLPCNN TransformersRAG
Data Engineering
ETL/ELTData Pipelines Data ModelingPreprocessing
MLOps & Deployment
Model EvaluationReal-time Inference VersioningGradio
Tools & Platforms
OpenCVGoogle Earth Engine Hugging FaceFAISS BitsAndBytes
Visualisation & Core CS
MatplotlibSeaborn DSADBMS OOP
Work

Featured Projects

NLP · RAG
Hallucination Detection using RAG
↑ 25–30% reliability improvement
Lightweight RAG-based framework using Phi-2, FAISS indexing, DeBERTa NLI cross-encoder, and HaluEval benchmarks. AND/OR-gate detection logic with 4-bit quantization for GPU-efficient inference.
PythonNLP TransformersRAGBitsAndBytes
MLOps · Digital Twin
AI-Driven Predictive Maintenance
↑ 20–25% operational efficiency gain
Predictive maintenance system using Digital Twin architecture to simulate real-time industrial environments and forecast equipment failures via optimised Python and SQL data pipelines.
PythonML SQLDigital Twin
IEEE 2024
SnapGesture — Real-Time Gesture Recognition
↓ 15–20% inference latency
CNN-powered real-time hand gesture recognition from live video streams. 9-layer CNN architecture with OpenCV pipeline for seamless, low-latency human-computer interaction.
PythonCNN OpenCVComputer Vision
SIH 2024
SAR Satellite Change Detection
↓ 15–20% false detection rate
Geospatial AI pipeline for automated change detection using SAR satellite imagery. Integrated Google Earth Engine with optimised computer vision for high-accuracy anomaly mapping.
PythonGeospatial AI Google Earth EngineCV
ML · Gradio
Stress Analysis with Interactive UI
↓ 15% response time
ML-based stress analysis system with an interactive Gradio interface for real-time prediction visualisation. Enables accessible, browser-based inference without server setup.
PythonMLGradio
IEEE Access
Uncertainty Quantification via MC Dropout
Confidence-aware predictions on CIFAR-10
Confidence-aware uncertainty quantification framework using Monte Carlo Dropout on CIFAR-10. Enables interpretable, reliable deep learning with calibrated prediction confidence scores.
PythonPyTorch Deep LearningBayesian ML
Research

Publications

2026
Hybrid Networks for Traffic Flow Detection and Road Accident Severity Analysis
IEEE · Accepted
2025
Distributed Training of Neural Networks in Smart Manufacturing Systems — Book Chapter in Distributed Deep Learning & XAI in Industry 4.0
Springer
2025
BERT–BART Fusion Model for Abstractive Text Summarization
IEEE
2024
SnapGesture: CNN-Powered Real-Time Hand Gesture Recognition
IEEE
Leadership

Roles & Contributions

2024 – 2025
Joint Secretary
Artificial Intelligence Association · Kongu Engineering College
Coordinated academic and technical activities in AI/ML, organised events, sessions, and student-led initiatives to enhance peer engagement and knowledge sharing.
2023 – 2024
Newsletter Head
Artificial Intelligence Association · Kongu Engineering College
Led editorial management of the association's technical newsletter — curating AI-related content, coordinating with contributors and faculty, resulting in increased readership and engagement.
Get In Touch

Contact Me

I'm open to research collaborations, internship opportunities, and discussions about AI/ML projects. Feel free to reach out!

+91 76048 86087
ahalyarajan003@gmail.com
linkedin.com/in/ahalya-rajan
github.com/Ahalya003
Coimbatore, Tamil Nadu, India