Spriha Podder

Computer Science undergraduate with a passion for impactful AI. Published researcher (ASCE 2025), Interned at Deloitte, UCL London Summer School alumna, and developer of neuro-AI and mental health-focused models, with proven academic and research excellence with global exposure and top-tier communication skills. Experienced in ML, Explainable AI (XAI), MLOps, full-stack dev, and cloud tools. Seeking AI/ML and development roles with real-world impact.

• 4× Research Paper Publications - ASCE ICDT 2025
• 4× Dean's List Recognition for academic excellence - Manipal University Jaipur
• UCL Summer School - First Division

Senior Promotions Coordinator at IEEE WIE MUJ 

Human Resources Coordinator at IAESTE India LC MUJ 

Marketing Team Member at  ACM MUJ SIGBED Chapter 

Core Committee Member for organising GENESIS Technical Fest by  IEEE MUJ 

Core Committee Member for organising Melange 24' - International Cultural Evening

Core Committee Member for organising Linguathon - A German Teaching Event

Core Committee Member for organising NexTrack - an IOT-based Line Following Robot Racing Event

Calm Compass 

Jun 2025 - Jul 2025

Live Prototype : https://calmcompassfrontend.vercel.app

• AI-Powered Travel Companion for Neurodiverse Users
• Designed a full-stack web app integrating a Python-based ML model to suggest calming travel routes and personalised mental wellness tips
• Built with React (frontend), Flask (backend), and deployed via Vercel & Railway
• Enables emotion-aware interaction and location-based route recommendations for stress-free journeys

Xenvue – turning datasets into compelling narratives

Jul 2025 - Aug 2025

Github Repo : https://github.com/SprihaPodder/Xenvue

• A contextual AI report generator that takes your raw numbers, stares deep into their soul, and spins up impressive, interactive reports
• Frontend React (Create React App, Vite, icons, and chart wizardry), modern UI, user authentication
• Backend: Python (Flask and FastAPI), scikit-learn, Perplexity API, web APIs (because I like a challenge)
• Dev tools: Git, npm, pip, neat repo hygiene
• Extras: CSV file uploads, stylish charts, authentication that actually works, all wrapped up in security and best-practices

A Lightweight Multi-Task Deep Learning Framework for Simultaneous Classification of Alzheimer's Disease Stages and Parkinson's Disease with Explainable AI

Manipal University Jaipur, Jan 2025 - May 2025 

Colab Notebook : https://colab.research.google.com/drive/1BvEpaMZ7or5P6f10tzy_8542M-NN9wSY?usp=share_link

• Developed a lightweight multi-task deep learning model to classify Alzheimer's, Parkinson's, and Control subjects from brain MRIs, with AD stage sub-classification
• Used a multi-task CNN with two heads and a modified combined loss function for simultaneous classification
• Integrated ResNet18-based CNN with clinical data for improved accuracy, and added LIME, Grad-CAM, and SHAP for explainability Enabled unified, interpretable diagnosis of neurodegenerative diseases

PhotonFreak - because your EV deserves smarter charging than your phone

Jun 2025 - Jul 2025

Github Repo : https://github.com/SprihaPodder/PhotonFreak

• Built predictive models for battery SOH (linear regression) and solar output (polynomial regression)
• Integrated FastAPI backend with React frontend for real-time prediction via REST APIs
• Deployed scalable, low-latency POC simulating intelligent EV charging based on weather and usage data

Emotion Aware Mental Health Support System using Game Theoretic Logic and Artificial Intelligence

May 2025 - June 2025

Colab Notebook : https://colab.research.google.com/drive/1OvWQJYTimcqALK1dJqMhz0QsdSsKeeeP?usp=share_link

• Designed an AI-powered, emotion-aware mental health support system that detects user emotions through NLP and responds strategically using game theory and modal logic
• The system models empathetic, adaptive conversations to promote cognitive well-being, combining affective computing with decision-theoretic reasoning for personalized support

ASCE International Conference on Transportation and Development, Glendale, Arizona, US

Jun 2025

Co-authored papers : 

A Diffusion-based model for price adjustment dynamics in delayed road construction projects

• Lead modeling and simulation using Python
• Implemented RBF (Radial Basis Function) networks to simulate delayed price diffusion across time segments in real-world contracts

Cost variation prediction in road construction: A Fuzzy Logic and Clustering Approach

• Co-designed fuzzy inference system for classifying key cost drivers
• Applied K-means clustering to segment project-level deviations, and optimise budget forecasting

Leveraging Radial Basis Function Regression and Bootstrapping for Price Adjustment in Road Contract

• Built regression framework using RBF with bootstrapped sampling
• Achieved robust variance-aware predictions of contract price adjustments over project duration.

Hazard-Based Model for Setting Financial Milestones in Road Construction Contracts

• Developed a survival analysis-based hazard model using Python and Excel
• Predicted time-to-milestone risks and optimized fund disbursement schedules