Skip to content
View Stokesy-dev's full-sized avatar
:electron:
Python
:electron:
Python
9 followers · 11 following

Achievements

Achievement: Quickdraw

Achievements

Achievement: Quickdraw

Highlights

  • Pro

Block or report Stokesy-dev

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Stokesy-dev/ReadMe.md

Hi, I'm Soham Warad 👋

B.Tech CS (AI & Data Science) | Building intelligent systems that solve real problems

I'm a software engineer and ML practitioner focused on shipping production-ready solutions. I build end-to-end systems — from data pipelines and ML models to scalable backend APIs and interactive frontends.

🔧 What I Build

Full-Stack ML Systems

  • RAG pipelines with vector search (FAISS, semantic retrieval)
  • Real-time systems using PostgreSQL LISTEN/NOTIFY and SSE
  • NLP classifiers with explainable AI (SHAP, attention visualization)

Data Science & Analytics

  • Time-series forecasting (Prophet, ARIMA, ML ensembles)
  • Anomaly detection systems (Isolation Forest, Z-score monitoring)
  • Graph ML (PyTorch Geometric, GCN, community detection)

Production Engineering

  • RESTful APIs and database design
  • System architecture for scale and reliability
  • Test-driven development and CI/CD

💻 Tech Stack

Languages: Python, JavaScript/TypeScript, SQL, R
ML/DS: PyTorch, scikit-learn, spaCy, Hugging Face, Prophet
Backend: Node.js, Express, PostgreSQL, Redis
Tools: Git, Docker, FAISS, Streamlit, Jupyter

🚀 Featured Projects

RefBot RAG

Football rules Q&A bot using RAG with FAISS vector search and Groq's Llama 3.1. Features page-level citation tracking and structure-aware chunking.

Financial KPI Dashboard

Enterprise forecasting platform with Prophet, ARIMA, and Random Forest. Built with TDD, includes SHAP explainability and multivariate anomaly detection.

Compliance NLP Classifier

Regulatory document classifier with DistilBERT, NER extraction, and conversational RAG. Features neural attention heatmaps for explainability.

Real-Time Order Tracking

Push-based order updates using PostgreSQL LISTEN/NOTIFY and Server-Sent Events. Zero polling, minimal latency. Live Demo

Character Network Analysis

Graph ML on Star Wars interactions using PyTorch Geometric. Implements GCN and GraphSAGE for node classification and centrality analysis.

📫 Let's Connect

🎯 Open to internship opportunities in Software Engineering, Data Science, and ML Engineering

Building at the intersection of software engineering and machine learning.

Pinned Loading

  1. Atypical-Assignment Atypical-Assignment Public

    Real-time order tracking system using PostgreSQL LISTEN/NOTIFY and Server-Sent Events for push updates without client polling

    JavaScript

  2. Compliance Compliance Public

    NLP classifier for regulatory compliance documents with explainable AI, NER extraction, and RAG-based conversational Q&A using DistilBERT and FAISS

    Python

  3. cp-coach cp-coach Public

    An AI-powered competitive programming coach that analyzes your Codeforces weaknesses and dynamically adjusts problem difficulty.

    TypeScript

  4. Financial-kpi Financial-kpi Public

    Enterprise financial forecasting dashboard with Prophet, ARIMA, and Random Forest. Features SHAP explainability and multivariate anomaly detection using Isolation Forest

    Python

  5. intelligent-ocr-digitalization intelligent-ocr-digitalization Public

    End-to-end Intelligent OCR system built with OpenCV, Tesseract, spaCy, and scikit-learn. Features robust photo preprocessing, named entity recognition, and mathematical text summarization without r…

    Python

  6. refbot-rag refbot-rag Public

    RAG-powered Q&A bot for football rules using IFAB Laws of the Game, FAISS vector search, and Groq's Llama 3.1 with citation tracking

    Python