Gesture (Accessibility Gestures)
Cross‑platform hand‑gesture control for desktop workflows (volume, workspace switching). 95% gesture accuracy with custom CNNs + Mediapipe.
Electron.jsMediapipeTensorFlowPyAutoGUI
Open to 2025–26 SWE internships
Bhinu Santosh
Aspiring Software Engineer • CS + DS @ UW–Madison
I build fast, reliable software with a soft spot for AI, clean backends, and UIs that don't fight you. I’m currently focused on backend systems, data pipelines, and machine learning integrations. (Yes, I commit code before coffee. Questionable life choice; great throughput.)
About
I like turning ambiguous problems into shipped products. Recent work: authentication, rate‑limiting, and data/AI plumbing on Azure. I aim for thoughtful UX, clean architecture, and feedback loops faster than a comet on espresso.
Skills
Python
TypeScript
React
Next.js
Node.js
FastAPI
Flask
Azure Functions
Azure SQL- Blob Storage
- Cosmos DB
PostgreSQL
Redis
Docker
CI/CD
JWT/Auth
PyTorch
TensorFlow
Projects
HackSherpa
Scraped 1,000+ DevPost projects to suggest tech stacks. Auto‑generates README + pitch, cutting planning time ~40%.
Next.jsSupabaseFlaskPythonTypeScript
A Jam Records APIs
Backend APIs on Azure Functions with secured data pipelines via Blob Storage + Azure SQL. Used APIM + logging to reduce latency and improve DX.
PythonAzure FunctionsAzure SQLAPIM
Experience
Software Engineering Intern
A Jam Records
- Developed and deployed Python/Azure Functions APIs, accelerating integration for a music‑tech startup.
- Built secure pipelines with Blob Storage + Azure SQL to move data between services and AI models.
- Streamlined arch + cut latency with Azure API Management + logging tooling.
Data Analyst Intern
Diyar United Company
- Drove +5% monthly revenue with interactive Power BI dashboards (churn, acquisition, revenue).
- Designed star schema enabling segmentation; contributed to +12% client acquisition.
- Automated 4 recurring ETL/reporting workflows, saving ~15 hrs/week.
Undergraduate Researcher
GLUE Lab, UW–Madison
- Mapped land‑use transitions in Amazon/Cerrado; improved geospatial accuracy for partners.
- Analyzed 20+ years of imagery & supply chains with PyTorch; linked trends to cattle/soy expansion.
- Reduced error in livestock classification by ~20% with optimized detection models.
Contact
Prefer email? Reach me at bhinusantosh@gmail.com. I’m always up for backend chats, ML rabbit holes, or debating tabs vs spaces (we all know the correct answer is tabs, right?).