Available for SDE & ML roles · 2027

Hi, I'mAllwin Joseph

AI & Data Science undergraduate building intelligent systems machine learning, generative AI and scalable backends that turn ideas into measurable impact.

PythonMachine LearningNLPFastAPISQLData Analytics
Portrait of Allwin Joseph

96.7% precision

IDS on NSL-KDD

2× Hackathon Winner

SIH Top 150 (internal)

About

Engineer, builder, perpetual learner

I'm an AI & Data Science undergraduate at Rajalakshmi Engineering College, Chennai. My work lives at the intersection of machine learning, generative AI and backend engineeringturning research-grade ideas into production-ready systems.

I love hackathons, hard problems, and shipping things that move real metrics. Beyond the keyboard I lead teams, speak publicly, and translate complex models into clear, persuasive stories.

Problem SolverTeam LeadPublic SpeakerHackathon Winner

3+

Featured Projects

96.7%

Best Model Precision

Hackathon Wins

Top 150

SIH Internal

Education

Academic foundation

2023 — 2027 · Expected

Rajalakshmi Engineering College

B.Tech — Artificial Intelligence & Data Science

Chennai, Tamil Nadu
AIData ScienceDeep LearningNLPStatistics

Experience

Where I've shipped

Real teams, real datasets, real outcomes.

Internship

Data Science Intern

Soil Steps · Student-led startup

Contributed to data-driven initiatives across the analytics pipeline — from raw data collection through model development to executive-ready insight.

  • Data collection & preprocessing
  • Exploratory data analysis (EDA)
  • ML model development
  • Business insight generation
  • Data visualization & reporting

Selected Work

Projects & case studies

A short tour of recent builds across ML, NLP and analytics.

Feb 2026

Security Analytics

Network Anomaly Detection System

Robust ML-based Intrusion Detection System achieving 96.7% precision on the NSL-KDD dataset, with live packet simulation via Scapy.

  • 96.7% precision on NSL-KDD
  • Feature engineering on network traffic
  • Random Forest, XGBoost, Logistic Regression
  • Reduced FP rate via tuning & CV
PythonScikit-LearnXGBoostPandas
Nov 2025

Natural Language Processing

Fake News Identification via Bi-LSTM

Context-aware news classifier built on a Bidirectional LSTM with custom embeddings, trained over 20,000+ articles.

  • Bi-LSTM architecture for context
  • 20,000+ articles preprocessed with NLTK
  • Custom embedding layers
  • Factual / biased / satirical classification
PythonTensorFlowNLTKKeras
Aug 2025

Ed-Tech Analytics

Student Performance & Risk Predictor

Classification system identifying at-risk students, with actionable intervention strategies to improve retention.

  • At-risk student classification
  • Preprocessing, encoding, normalization
  • Intervention strategy generation
  • Statistical performance visualizations
PythonScikit-LearnPandasMatplotlib

Skills

Tooling & expertise

The stack I reach for to ship intelligent, reliable systems.

Languages

  • Python92%
  • SQL / MySQL85%
  • Java (Basic)55%

AI / ML

  • Machine Learning90%
  • Deep Learning82%
  • NLP85%
  • Generative AI78%

Backend

  • FastAPI85%
  • REST APIs88%
  • Auth Systems75%

Tools & Data

  • Pandas / NumPy90%
  • Scikit-Learn88%
  • Power BI78%
  • Docker / Git75%

Recognition

Achievements & wins

A few moments where ideas met execution under pressure.

Winner — Conocithon

1st Place

Park Engineering College, Coimbatore

Winner — Codeathon

1st Place

Prathyusha Engineering College, Chennai

Smart India Hackathon

SIH 2024

Top 150 Teams Shortlisted Nationwide

Team Leadership

Leadership

Led multi-disciplinary teams across hackathons & projects

Contact

Let's build something

Open to internships, collaborations and SDE / ML roles.

Best way to get a reply is a quick line about what you're building.

allwinjoseph.2605@gmail.com