Open to AI & software engineering roles

Joshua JJ Wonder

Turning AI research into practical products.

I'm an ECE graduate and edge-AI/full-stack developer building intelligent systems across generative AI, computer vision, signal processing, APIs, and mobile applications.

01IEEE-publishedAIMLA 2026
02Edge deployedLocal Llama-3 8B
03Cross-disciplinaryAI · DSP · Product
04Production mindedAPIs · Tests · Deployment

Selected work / 2023—2026

Systems with a reason to exist.

Research depth, product thinking, and engineering execution—shown through the problems I chose and the systems I shipped.

02Deep learning · Computer vision
0.208 BLEU-4

Image Caption Generator

An encoder-decoder model with Bahdanau attention, ResNet-50 spatial features, and a Gradio interface for turning images into natural-language captions.

  • PyTorch
  • ResNet-50
  • LSTM
  • Gradio
View project
03Full-stack product
Multi-role SaaS

ConvoManage

A multi-role conference platform for organisers, speakers, and attendees, with authentication, scheduling, real-time updates, and payment-ready workflows.

  • React
  • TypeScript
  • Supabase
  • PostgreSQL
View project
04Backend engineering
100% coverage

Smart Expense API

A clean FastAPI service with strict Pydantic validation, interactive OpenAPI documentation, contract tests, and 100% line-coverage enforcement.

  • Python
  • FastAPI
  • Pydantic
  • Pytest
View project
More experiments: PulseMix, Bible App, license-plate detection, and credit calculators

About / The crossover

Engineering, from the signal to the screen.

My ECE background taught me to think in systems: understand the signal, model the constraints, measure the result, and make the whole thing work.

That mindset now shapes how I build software. I'm most engaged when a project crosses boundaries—AI with product design, vision with edge hardware, or research with an interface that people can actually use.

Download full résumé
ENGINEERING TOOLKIT12 MODULES
  • 01Python
  • 02TypeScript
  • 03C++
  • 04React
  • 05React Native
  • 06FastAPI
  • 07PyTorch
  • 08OpenCV
  • 09PostgreSQL
  • 10Docker
  • 11Linux
  • 12Git

Research note / AIMLA 2026

Measure what matters. State what remains unknown.

ρ0.484

Spearman correlation between the DSP engagement score and quiz performance in a 21-participant pilot.

EduSync models scrolling as a discrete-time signal, then combines FIR smoothing, zero-crossing rate, frequency analysis, and vision-based attention estimates into interpretable learning analytics.

The pilot produced promising positive associations, but its small sample and multiple-comparison limits matter. I present the work as a measured starting point—not a finished scientific claim.

Publication details

Experience

Learning by shipping.

DEC 2024 — JAN 2025

Web Development Intern

Innovate Intern · Remote

Built full-stack workflows for ConvoManage, improved SQL-backed data handling, and collaborated on debugging, deployment, and version control.

MAY 2024 — JUN 2024

AI & Machine Learning Intern

Innovate Intern · Remote

Developed a YOLO-based license-plate detection pipeline spanning preprocessing, annotation, training, tuning, and real-world edge-case testing.

2022 — 2026

B.Tech, Electronics & Communication

Karunya Institute of Technology and Sciences

Built a foundation in signal processing, electronics, programming, and systems engineering—then applied it to AI and product development.

Now receiving signals

Have an ambitious problem?

I'm looking for engineering teams where curiosity, measurement, and thoughtful execution all matter.

Start a conversationwonderjj2017@gmail.com