I train models, deploy infra, write code, and sometimes make music — not always in that order.
aakash = {
"location" : "Bangalore, India 🇮🇳",
"currently" : "Associate Systems Engineer @ IBM",
"prev" : ["Averlon Enterprise Solutions", "Tata Consulting Engineers Ltd."],
"focus" : ["Machine Learning", "Fullstack Dev", "DevOps", "Cloud"],
"homelab" : "Proxmox on a Lenovo Yoga — because why rent servers?",
"music" : ["Piano 🎹", "Guitar 🎸", "Ear training nerd"],
"code" : ["Python 🐍", "JavaScript ⚡", "SQL", "C++"],
"speaks" : ["English", "Hindi", "Marathi"],
"fun_fact" : "I once compared 3 forecasting models to predict energy demand. Random Forest won. It always wins.",
}ML / AI
Fullstack
Databases
DevOps / Infra
|
Personal finance platform covering banking, investments, insurance, and expense tracking — built from the ground up with a clean design-first approach. |
Running Proxmox on a repurposed Lenovo Yoga. Hosts NAS, personal apps, and project deployments. My own little on-prem cloud. |
| Project | Description | Stack |
|---|---|---|
| 💰 Senary Capital | Personal finance app: banking, investments, insurance & expense tracking with a unified dashboard | Next.js, PostgreSQL, Prisma ORM, Tailwind CSS |
| 🎵 EchoDeck | Collaborative music streaming SaaS: shared queues, sync playback, role-based access control, full auth system | Next.js, PostgreSQL, Prisma ORM, Tailwind CSS, PHP |
| 🧘 ZenKraft | AI-powered yoga assistant with real-time pose correction, ML-driven sequence progression and user tracking | Python, React.js, Node.js, MongoDB, OpenCV, MediaPipe |
| ⚡ Energy Demand Forecasting | Predictive models for electricity demand, compared RF, XGBoost & ARIMA, evaluated on RMSE, MAE, MAPE | Python, Scikit-learn, XGBoost, Pandas |
| 🏠 Personal Homelab | Distributed homelab on repurposed hardware: reverse proxy, NAS, password manager, media server, Tailscale mesh | Proxmox, Docker, Linux, Tailscale |
When I'm not pushing commits or tuning hyperparameters, I'm at the piano or guitar, recreating songs by ear, performing, and occasionally competing. Music and ML aren't that different. Both are about finding patterns in noise.
"Ship it. Learn from it. Rebuild it better."
⭐ If any of my work helped you, a star goes a long way!