Skip to content
View SWAPI03's full-sized avatar
🎯
Focusing
🎯
Focusing
5 followers · 4 following

Achievements

Achievement: Quickdraw

Achievements

Achievement: Quickdraw

Block or report SWAPI03

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.

Content in all repositories owned by your account will be closed.
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
SWAPI03/README.md
Swapnil Yadav: ML Engineer, Data Scientist, AI Builder

Typing intro

LinkedIn Email Kaggle LeetCode


About me

class Swapnil:
    based_in  = "Lucknow, India"
    studying  = "B.Tech CSE (AI & DS) @ IIIT Manipur, 2023-27, CGPA 8.31"
    interned  = ["ML Intern @ FlyRank AI", "Data Science Intern @ DecodeLabs"]
    focus     = ["explainable AI", "computer vision", "LLM apps", "MLOps"]
    building  = "CVE Copilot, an LLM assistant for vulnerability triage"
    learning  = ["RAG", "fine-tuning", "agents"]
    open_to   = ["SWE / ML internships", "open source", "interesting ML problems"]

I like ML that can explain itself: Grad-CAM heatmaps, SHAP values, evals with real numbers behind them. Most of my projects end in an app or a dashboard, not just a notebook.

Featured work

Featured projects. CH.01 CVE Copilot: LLM assistant for CVE triage, 100% within one severity level, about 0.07 INR per CVE. CH.02 Chest X-Ray Pneumonia Detection: DenseNet121 + Grad-CAM in TensorFlow/Keras, 0.94 test AUC, 99.7% pneumonia recall, 84.6% accuracy. CH.03 Reliability-Aware AutoML: trust-weighted training tested on 60K+ rows across 4 datasets. CH.04 Telco Churn: 0.84 ROC-AUC, churn recall raised from 0.51 to 0.78.

READ ▸ CH.01 CVE Copilot · CH.02 Chest X-Ray · CH.03 Reliability AutoML · CH.04 Telco Churn

More chapters
Project What it does Stack
Fake News Detection TF-IDF + Linear SVM on 44K news articles, 99.6% accuracy scikit-learn NLTK
Walmart Sales Forecasting Weekly sales forecasts across stores and departments Pandas scikit-learn
Django Sentiment Analysis Sentiment classifier served through a Django web app Django NLTK
Pet Breed Classifier (in progress) ResNet transfer learning on 37 cat and dog breeds PyTorch torchvision
MLOps with DVC Versioned data and reproducible ML pipelines DVC Git

Achievements

Kaggle competition bronze medal: Biohub Cell Tracking During Development, 287th of 3,947 teams, top 8%

  • Kaggle Competition Bronze Medal: Biohub Cell Tracking, 287th of 3,947 teams (top 8%)
  • NPTEL Business Intelligence & Analytics: Elite + Top 2%, scored 97%
  • NPTEL Introduction to Information Retrieval: Elite + Top 2%, scored 93%
  • 250+ LeetCode problems solved, 100+ of them medium or hard
  • Merged contributions to crewAI (59k★) and Composio (30k★)

Open source

I fix bugs in the AI tooling I use. Recent work in crewAI, LiteLLM, Composio and redis-py:

Project Pull request Status
crewAI Raise ValueError instead of a bare raise in get_uploader
Composio Fix .jpg resolving to non-standard image/jpg
xblock-ai-evaluation Fix language field using Scope= instead of scope=
LiteLLM Label .webm as audio/webm in audio transcription
crewAI Make orphaned-shard cleanup work on Windows
crewAI Persist evaluation results as UTF-8
redis-py Fix JSON.set_path key derivation to strip only the extension

See all my pull requests →

Experience

Machine Learning Intern · FlyRank AI  Jul 2026 - Sep 2026
Applied Search Intelligence: used ML to study how pages rank and get discovered on Google, working with FlyRank's anonymized search data (up to ~79M rows, queried through DuckDB). Work repo

Data Science Intern · DecodeLabs  Jun 2026 - Jul 2026
Built end-to-end ML pipelines covering EDA, feature engineering, SMOTE, PCA and K-Means, plus a fraud detection pipeline with ROC-AUC based model selection. Task 1 · Task 2

Toolkit

Python, C++, C, Java, PyTorch, scikit-learn, Django, MySQL, SQLite, Redis, HTML, CSS, Git, GitHub Actions, VS Code

Area Tools
ML / DL PyTorch scikit-learn XGBoost SHAP Grad-CAM
LLMs OpenAI API Hugging Face crewAI LiteLLM
Data Pandas NumPy DuckDB NetworkX Matplotlib Seaborn
Ship it Streamlit Django DVC Jupyter GitHub Actions

Stats

GitHub streak LeetCode stats

Contribution graph being eaten by a snake

Pinned Loading

  1. reliability-automl reliability-automl Public

    AutoML pipeline that scores data quality across 5 dimensions, propagates row level trust through a similarity graph, and trains models with reliability weighted samples — built with Streamlit, scik…

    Python 2

  2. cve-copilot cve-copilot Public

    Jupyter Notebook

  3. Chest_Xray_Disease_Detection Chest_Xray_Disease_Detection Public

    Deep Learning-based Chest X-ray Disease Detection using PyTorch with Grad-CAM visualization, comprehensive model evaluation, and explainable AI techniques.

    Jupyter Notebook 2

  4. flyrank-ai-internship flyrank-ai-internship Public

    Jupyter Notebook 2

  5. django-sentiment-analysis django-sentiment-analysis Public

    Python 2

  6. Fake_News_Detection_NLP Fake_News_Detection_NLP Public

    End to end NLP pipeline for fake news detection using TF-IDF, Logistic Regression, and Linear SVM with comprehensive evaluation and visualization.

    Python 2