Hi, my name is

Jonathan.

> currently.building: smarter tools for real-world problems

I build AI-powered tools and systems
that solve real-world problems.

Computer Science student focused on AI, cybersecurity, and scalable systems.

Built and deployed 4 real-world systems across AI, cybersecurity, and large-scale data processing.

View My Work View Resume ↗ Connect

scroll

01. > featured.projects

Featured Project

Ludex — Hybrid Recommendation System

A hybrid recommendation engine fusing TF-IDF content modeling with implicit-feedback collaborative filtering (ALS). Serves personalized game recommendations via anchor-based profiling, score-normalized fusion, and MMR-based re-ranking — handling cold-start users out of the box.

Built with: Python, Scikit-learn, Implicit ALS, Steam API

Machine Learning
Recommendation Systems
Python

Featured Project

File Converter — Document Processing Engine

A deterministic two-pass document conversion engine — structural analysis separated from rendering. Handles paragraph reconstruction, list detection, heading inference, and conservative table extraction with fully explainable outputs. No OCR, no ML, fully deterministic.

Evolving into a full document processing suite with multi-format conversion and PDF compression.

Built with: Python, Two-Pass Architecture

Python
Document Processing
Systems Design

Featured Project

WebGuardian — Phishing Detection System

Multimodal phishing detection combining Char-CNN + LSTM (URL analysis) and MobileNetV2 (visual analysis) via late fusion. Achieves 98.73% accuracy on 3,000 real-world webpages with ~35ms inference — deployed as a production Chrome extension for real-time, on-device threat detection.

Built with: Python, Char-CNN + LSTM, MobileNetV2, Chrome Extension

Deep Learning
Cybersecurity
Computer Vision

Featured Project

Synthetic Data Object Detection — Detection Pipeline

End-to-end object detection trained on zero real-world images — entirely synthetic data generated in Blender with randomized 3D scenes. YOLOv8 Nano on 640×640 renders achieves 0.997 precision, 1.000 recall, and 0.995 mAP@50, proving synthetic-only training can match real-data performance.

Built with: Python, Blender, YOLOv8, Ultralytics

Computer Vision
Synthetic Data
Deep Learning

Ludex personalized game recommendations interface

PDF → Parse → Structure → Render → DOCX

Synthetic object detection with bounding boxes

Want to explore more?

View All Projects on GitHub →

02. > skills.list
[ system ] [ visual ]

[ CORE ] Languages

Python (core) → used for backend systems, automation, and machine learning pipelines
Java (oop) → object-oriented programming for scalable application logic and systems
C / C++ (systems) → low-level programming for performance-critical and system-level tasks
Dart (ui) → used for building cross-platform UI with Flutter
Flask (api) → lightweight backend framework for building REST APIs
Flutter (mobile) → cross-platform mobile app development with a single codebase

[ LOADED ] Libraries & Tools

PyTorch (training) → deep learning framework for model training and experimentation
TensorFlow (models) → production-ready ML framework for building and deploying models
OpenCV (vision) → computer vision library for image processing and detection tasks
NumPy (compute) → numerical computing and array operations for ML/data workloads
Pandas (data) → data manipulation and preprocessing for structured datasets

[ ACTIVE ] AI / Machine Learning

Machine Learning (pipelines) → end-to-end ML pipelines including preprocessing, training, and evaluation
Deep Learning (cnn) → neural network architectures for high-dimensional data like images
Computer Vision (analysis) → extracting insights from images and video using ML techniques
Recommender Systems (matching) → building systems that suggest relevant content based on patterns
YOLO (real-time) → real-time object detection optimized for speed and accuracy

[ SECURE ] Cybersecurity

Threat Detection (real-time) → identifying malicious patterns and anomalies in real-time systems
Phishing Analysis (nlp/cv) → detecting phishing attacks using NLP and visual analysis
Network Security (monitoring) → monitoring and securing network traffic and system behavior

[ RENDER ] Graphics / Simulation

Blender (bpy) → scripting and automation for 3D graphics and simulations
Synthetic Data (generation) → generating artificial datasets for training ML models

[ ONLINE ] Systems Engineering

System Architecture (design) → designing scalable and modular system structures
Pipeline Design (etl) → building data pipelines for processing and transformation workflows
Performance Optimization (optimization) → improving efficiency and speed of systems and algorithms
GPU Computing (cuda) → orchestrating cloud infrastructure and scalable container architectures

03. > about.me

I'm a 3rd-year Computer Science student at IIIT Pune who builds real tools - not just coursework. My focus areas are AI, cybersecurity, and systems design, and I gravitate toward projects that solve practical, tangible problems.

Whether it's building a multimodal phishing detector as a browser extension, generating synthetic training data in Blender, or engineering a deterministic document converter - I focus on software that works in the real world.

I'm always working on something new. Currently leveling up and looking for opportunities to build at scale.

developer.js

const developer = {
  current: "B.Tech CSE, IIIT Pune",
  year: "3rd Year",
  location: "Pune, India / Muscat, Oman",
  focus: ["AI", "Systems", "Security"],
  status: "open to internships"
};

Jonathan.

I build AI-powered tools and systems
that solve real-world problems.