About Work Research Experience Recognition Contact
Open to Internships & Opportunities

Soumili Saha

AI & Machine Learning / Software Engineering / Research & Systems

I build ML systems that solve real problems — disaster risk prediction, environmental health AI, and data-driven decision tools. Currently a 2nd-year CSE student specialising in AI & ML with a minor in Cybersecurity.

Python TensorFlow XGBoost RAG Flask Scikit-learn CNN
Soumili Saha

Building systems
for real-world impact.

I'm a Computer Science undergraduate at IEM Kolkata, majoring in AI & Machine Learning with a minor in Cybersecurity. My focus is applied ML — building end-to-end systems that move from raw data to deployed, usable tools.

My work deals with high-stakes problems: predicting natural disasters before they happen, detecting disease risk from environmental images, classifying health risk factors in under-resourced areas. The engineering has to be reliable because the problems are real.

Outside of project work I study algorithms, earn certifications, compete in hackathons, and explore research-driven problem solving at the intersection of AI, public health, and environmental science.

I'm a Computer Science undergraduate at IEM Kolkata specializing in AI & Machine Learning with a minor in Cybersecurity.

I build end-to-end ML systems focused on disaster prediction, environmental risk analysis, and applied AI.

My interests lie at the intersection of AI, public health, environmental science, and research-driven problem solving.

IEM Kolkata
B.Tech CSE (AI & ML) · 2024–2028 · CGPA 8.79
AI · ML · Cybersecurity
Major Specialisation & Minor
IBM · BeeSkilled
Active ML & AI Internships · AICTE Certified
Technical Skills
Core ML
Python TensorFlow XGBoost Scikit-learn Keras CNN
AI Systems
RAG Pipelines Llama 3.2 EfficientNetB0 MobileNetV2 PCA K-Means
Data & APIs
Pandas NumPy Matplotlib Seaborn REST APIs OpenWeather API
Engineering
Flask Streamlit Git Java SQL JavaScript HTML/CSS

Projects

2025–2026
disaster-risk-prediction.streamlit.app
Multi-Hazard Disaster Risk Prediction System Screenshot

Multi-Hazard Disaster Risk Prediction System

Real-time flood and heatwave risk assessment using machine learning models and live weather APIs.

K-Means + XGBoost Random Forest Multi-Hazard AI Streamlit Live Weather APIs
environment-health-system.local
Environmental Health Risk & Advisory System Screenshot

Environmental Health Risk & Advisory System

CNN + RAG system for detecting disease-associated environmental hazards and generating preventive health advisories.

EfficientNetB0 FAISS RAG LLaMA 3.2 Ollama Flask
cortex-arena.run.app
Cortex Arena: AI Crowd Orchestration System Screenshot

Cortex Arena: AI Crowd Orchestration System

AI-powered crowd management platform with congestion monitoring, route optimization, and attendee assistance.

FastAPI WebSockets Gemini Cloud Run
aidflow-ai.vercel.app
AidFlow: Agentic Volunteer Intelligence System Screenshot

AidFlow: Agentic Volunteer Intelligence System

Multi-agent disaster response system that transforms reports into actionable volunteer coordination plans.

Gemini Multi-Agent AI Disaster Response Google Cloud

Academic Work

Applied research projects, conference submissions, and ongoing technical investigations

Applied AI Research Submitted

An AI-Powered Intelligent Advisory System using CNN and RAG Based Framework for Detecting Health Risks from Environmental Factors

Research on detecting disease-associated environmental hazards from images using CNNs and generating context-aware preventive health advisories through a Retrieval-Augmented Generation (RAG) pipeline.

Conference Review Pending
Computer Vision NLP / RAG Public Health
ML & Decision Support Submitted

AI-Driven Risk Analysis of Floods and Heatwaves in Real-Time Environments

Research on real-time flood and heatwave risk assessment using machine learning, live meteorological APIs, and decision-support frameworks for localized hazard monitoring.

Manuscript Submitted
Ensemble ML Disaster Mgmt. Real-time Data
Security Analytics In Progress

Intelligent Network Anomaly Detection Framework

Research and development of machine learning techniques for multi-source anomaly detection, threat monitoring, and real-time cybersecurity analytics using interactive dashboard systems.

Research & Development
Cybersecurity Anomaly Detection Dashboard Analytics

Experience

May 2026 – Present
Current AICTE Remote

BeeSkilled

AI & ML Intern

Building predictive and clustering models using Python and Scikit-learn on healthcare and real-estate datasets. Focused on preprocessing, model selection, and evaluation.

Python Scikit-learn Pandas Model Evaluation
Jan 2026 – Feb 2026
AICTE Remote

IBM SkillsBuild × Edunet

AI & ML Intern

Completed AI/ML training covering data preprocessing, model evaluation, and model development. Built and deployed a disaster risk prediction application using Streamlit.

Python XGBoost Streamlit End-to-end ML

Achievements

Featured · NPTEL

Elite + Gold + Topper — Programming in Java

Achieved Elite + Gold certification with a perfect 100% score across all assessments, ranking among the top performers nationally.

Top 1% Elite + Gold NPTEL 2026
2025 · NPTEL
Elite + Gold — The Joy of Computing Using Python

Achieved Elite + Gold certification with a 90% score in NPTEL's computational thinking and Python programming course.

2025 · HackerRank
5-Star Python Proficiency

Earned a 5-star Python rating through problem solving and coding assessments.

2025 · Smart India Hackathon
SIH 2025 Pre-Final Round

Ranked among the top 50 teams from IEM Kolkata in Smart India Hackathon 2025, advancing to the pre-final round with a team-developed technical solution.

Retrieval Augmented Generation (RAG)
DeepLearning.AI
Probability & Statistics for Machine Learning & Data Science
DeepLearning.AI
Python for Data Science, AI & Development
IBM
AWS Cloud Practitioner Essentials
Amazon Web Services
Modern Data Warehouse Analytics in Microsoft Azure
Microsoft

Have an idea? Let's talk.

Focused on machine learning, applied AI, and software engineering. Whether it's research, internships, collaborative projects, or building something ambitious from scratch, I'm always open to thoughtful conversations.

Opens your default email client with the message pre-filled.