Soumili Saha

About

Computer science student specializing in AI/ML and full-stack development, with internship experience building AI platforms and machine learning pipelines. Works with modern web and ML stacks including Next.js, FastAPI, Python, Scikit-learn, TensorFlow, and XGBoost.

Experience

• Gen AI & Cloud Computing Intern
  BharatCares (CSRBOX Group)
  Jun 2026 – Jul 2026

  Shipped InterviewVerse AI, a full-stack interview platform supporting 3+ simulation modes spanning HR interviews, technical interviews, and project vivas. Built core platform functionality using Next.js, FastAPI, PostgreSQL, and Gemini API; separately completed AICTE x IBM SkillsBuild training in cloud computing fundamentals covering AWS and GCP.

• AI & ML Intern
  Edunet Foundation
  Jan 2026 – Feb 2026

  Developed machine learning pipelines using Python, Scikit-learn, and XGBoost for environmental and meteorological datasets. Integrated external weather APIs into a live Streamlit dashboard for multi-hazard risk assessment and visualization.

Education

• B.Tech in Computer Science Engineering (AI & ML)
  Institute of Engineering and Management
  Computer Science & Engineering (AI & ML) · Jul 2024 – Jun 2028
• Higher Secondary
  The Newtown School Kolkata
  Science · 2024
• Secondary
  The Newtown School Kolkata
  2022

Skills

Tools / apps / platforms

Projects

• Multi-Hazard Disaster Risk Prediction System
  Python, Scikit-learn, XGBoost, Streamlit

  Built a machine learning system for 2 hazards—flood and heatwave risk—using meteorological and environmental datasets. Combined 95–98% predictive accuracy with 2 real-time weather APIs in a Streamlit dashboard for location-based risk assessment.

• Environmental Health Risk and Advisory System
  Python, TensorFlow, Flask, Llama 3.2

  Built an AI system combining computer vision and LLM-based reasoning to analyze environmental images and generate public health insights. Benchmarked MobileNetV2 and EfficientNetB0 with a local Llama 3.2 RAG pipeline to generate contextual health recommendations.

• PaperYield – AI Research Gap Discovery Platform
  Next.js, FastAPI, PostgreSQL, RAG

  Built a full-stack platform processing 5–20 research papers, extracting 8+ structured attributes, and generating citation-grounded responses through a RAG pipeline. Implemented a retrieval workflow using structure-aware chunking, dense retrieval, BM25 keyword search, and cross-encoder reranking with page-level citations.

Courses & certifications

• The Joy of Computing Using Python · NPTEL · 2026
• Programming in Java · NPTEL · 2026
• AICTE x IBM SkillsBuild training in cloud computing fundamentals · AICTE x IBM SkillsBuild · 2026

🏆 Achievements & awards

• Research Paper accepted for presentation at ISICVA 2026 · 2026

  Accepted for presentation at the 2nd International Symposium on Innovations in Computer Vision and Applications.

• Smart India Hackathon 2025 · 2025

  Led a team in developing a technical solution and advancing through multiple evaluation rounds to the pre-final stage among the Top 50 teams.

• NPTEL: Programming in Java · 2026

  Elite + Gold + Top 1% (100%).

• NPTEL: The Joy of Computing Using Python · 2026

  Elite + Gold (90%).