Full-stack engineer working across Java and Python backends, machine learning, and on-chain systems. I like problems where messy real-world signals have to become clear, trustworthy decisions.
LatestShipped SolarPay at Monad Blitz Pune V3 — a solar market with real hardware and agents, built in a day and verified on Monad mainnet.01Selected work
Each project starts with the problem, then the decisions that shaped the system. Links go to the real code.
Rooftop solar that sells its surplus to the neighbour by itself — a real panel and Arduino meter, rule-based agents, and one Monad transaction per trade.
AI · Database intelligence
A semantic layer over an existing relational database, so business questions can be asked in plain language and answered from structured data.
Machine learning · Computer vision
A convolutional neural network that classifies leaf images as healthy, powdery mildew, or rust — served through a small Flask upload app.
Full-stack · Java · Python · Android
A banking REST API in Spring Boot, a gym operations platform in Flask, and a college app paired with an admin app on a shared Firebase backend.
02About
Om
Salunkhe
Software Engineer
I'm Om, a software engineer who learns by building things that have to work. My foundation is full-stack development — Java and Spring Boot, Python and Flask, relational data — and I've been pushing that foundation toward AI and data intelligence and toward blockchain systems where rules need to hold on their own.
Recently that meant building the gateway, agents, and web app for a solar energy market at a Monad hackathon, and a semantic layer that lets people query a database by asking questions in plain language.
The interesting part of a problem is usually the rule that can't break — a balance that can't go negative, a meter reading that can't be sold twice. I design around that first.
SolarPay keeps the LLM out of the money pathI'm comfortable moving from a serial port to a REST API to a React screen. Seeing the whole path is how I find the bugs that live between layers.
Arduino → Python gateway → Monad → web appWhether it's an agent choosing a seller or a model calling a leaf diseased, people should be able to see why. I log reasons, not just results — and report limitations plainly.
Plant CNN: results, honestlyMost data problems are context problems. Modelling what the data means — before asking an AI to use it — is what makes answers trustworthy.
Semantic layer over a relational database03Capabilities
No percentages. Each skill points at the project where it was used; areas I'm still actively learning are marked as such.
Java for structured backends and Android, Python for data, ML and systems glue.
Layered APIs, background jobs, and data models that enforce business rules.
Interfaces that make complex state legible — from dashboards to agent replays.
Turning raw signals and schemas into decisions people can trust and inspect.
On-chain settlement where rules must hold even when software misbehaves.
The unglamorous parts that keep systems testable and running.
04Milestones
Shipped a neighbour-to-neighbour solar market in one day with a real panel, rule-based agents, and a contract verified on Monad mainnet. Owned the gateway, agents, backend, and web app.
Case study: Built SolarPay at Monad Blitz Pune V3Trained a three-class leaf disease CNN with Keras and served it through a Flask upload app.
Case study: Plant disease classifierA layered Spring Boot REST API on MySQL, and a Flask operations platform with roles, attendance, and scheduled fee reminders.
Case study: Banking API and gym management platformTwo Android apps sharing a Firebase backend: one to publish notices and gallery images, one for students to read them alongside faculty listings and e-books.
Case study: College app + admin app05Open source
Every public project lives on GitHub — 6 public repositories. Judge the work by the source, not the summary.
Your rooftop solar sells surplus power to your neighbour, by itself. Real panel + Arduino meter, rule-based agents, one Monad tx pays seller, grid fee and green receipt. Live on Monad mainnet.
Spring Boot REST API for bank accounts — deposits, withdrawals, and account management on MySQL.
Flask platform for members, classes, payments, attendance, and automated fee reminders.
Keras CNN that classifies leaf images as healthy, powdery mildew, or rust, served with Flask.
Android admin app that publishes notices and gallery images to Firebase.
Android college app — notices, faculty by department, e-books, and gallery from Firebase.
Live from the GitHub API · refreshed hourly
View GitHub profile06Contact
I'm interested in engineering roles, freelance projects, and collaborations across full-stack, AI and data, and Web3. Tell me what you're working on.