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.

View Projects Get in Touch

PythonMachine LearningNLPFastAPISQLData Analytics

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

Featured Projects

96.7%

Best Model Precision

2×

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.

Apr 2026

MCP Based

Autonomous Security Investigation Platform

Autonomous security platform that leverages Model Context Protocol (MCP) and agentic reasoning to identify network anomalies and execute incident response.

Designed an autonomous security platform to identify network anomalies by leveraging system logs and real-time threat intelligence
Applied automated orchestration and agentic reasoning to correlate security data and initiate responses
Provided data-backed incident reports and standardized JSON ticketing to empower security admins

PythonMCPFastAPIOllamaReAct FrameworkScapySQLiteMachine LearningData AnalysisData Visualization

Code

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

Code

Apr 2026

AI / ML Integration

Local AI Workforce Copilot (RAG System)

An enterprise-grade local RAG pipeline integrating SQLite and multi-page PDFs with a Llama 3 FastAPI backend and Streamlit UI.

Developed an enterprise-grade RAG pipeline integrating SQLite records and multi-page PDFs into a unified vector search engine
Utilized LangChain and HuggingFace models to handle document chunking, embedding generation, and vector matching via FAISS
Connected retrieval to Llama 3 via Ollama/FastAPI, serving clean context-aware data to a Streamlit frontend

PythonLangChainFAISSFastAPIStreamlitOllamaSQLite

Code

Feb 2026

Security Analytics

Network Anomaly Detection System

Robust ML-based Intrusion Detection System (IDS) achieving 96.7% precision using the NSL-KDD dataset.

Performed feature engineering on network traffic data to improve detection accuracy of malicious patterns
Utilized Scapy for live packet simulation to identify novel attacks that bypass traditional rule-based systems
Built comparative models using Random Forest, XGBoost, and Logistic Regression for performance benchmarking

PythonScapyRandom ForestXGBoostLogistic Regression

Code

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%
Git40%

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

Testimonials

What people say

Feedback from mentors, teammates, and collaborators.

“

Allwin Joseph is a phenomenal problem solver who possesses a rare combination of sharp analytical thinking and rapid execution. He has an incredible ability to dissect complex, ambiguous problem statements and grasp core requirements with absolute precision. What truly sets Allwin apart is his efficiency—he consistently engineers elegant, reliable solutions well within tight, stipulated deadlines. With his agile mind and natural technical intuition, Allwin has immense potential to tackle any high-stakes challenge and deliver impactful results.

Manoj Kevin

Founder · Eva Beauty Lounge

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