GenAI Engineer Full Stack Dev ML Research Open to Collab Building in Public Bengaluru, India Agentic Systems RAG Pipelines Privacy Tools Trading Bots GenAI Engineer Full Stack Dev ML Research Open to Collab Building in Public Bengaluru, India Agentic Systems RAG Pipelines Privacy Tools Trading Bots
Developer Profile · 2025
HARSH
SINGHAL
Role
GenAI · Full Stack · ML Engineer
Status
Open to collab
Based in
Bengaluru, India
Building · Shipping · Breaking
What I obsess over
Agentic systems
Algorithmic trading
Privacy infrastructure
RAG pipelines
Dev tooling
Distributed systems
Manifesto
"Building AI systems that think, trade, defend, and review — so you don't have to."
I work at the intersection of applied AI and real-world systems. From agentic code reviewers to autonomous privacy enforcement platforms, the goal is the same: build something that actually does the job without hand-holding. I like problems where the solution is uncomfortable to build.
Research Highlight
Used satellite imagery + ML to predict urban flooding in Bangalore — processing geospatial raster data to generate flood-risk zone maps before events unfold.
Featured Work
3 Projects · Active
01
01
Code Review Automation · TypeScript · GenAI
Velocis
Active github.com/Velocis →
AI senior engineer embedded in your repo — it reads the diff, questions the logic, and enforces your team's standards without being asked.
Velocis plugs into your GitHub workflow as a virtual senior engineer. It understands your codebase context, reads every pull request fully, and returns actionable feedback — not surface-level linting warnings but real engineering critique: logic bugs, architectural drift, unhandled edge cases, naming that lies.
It's less a bot and more the dev on your team who actually reads the diff before commenting. It integrates via webhooks and GitHub Actions, meaning zero additional tooling in your flow. The feedback arrives in the PR thread like any other reviewer.
Auto
PR Review
Context
Aware
GitHub
Native
Zero
Extra Tooling
24/7
On-call
TypeScript Node.js GitHub API OpenAI Webhooks GitHub Actions
02
02
Next.js · Python · Groq · Binance
Quantis
Build, backtest & run crypto strategies — with an AI coach watching your every move.
Active
Full-stack crypto algorithmic trading simulator with a Monaco-powered in-browser strategy editor. Connect it to real Binance market data, run backtests across historical windows, and get Groq-powered AI coaching that explains exactly why your strategy is losing money.
The AI coach doesn't just analyse performance metrics — it reads your strategy code and identifies logical flaws, overfitting patterns, and risk exposure you probably didn't account for. Deployed on Cloudflare edge for sub-30ms global latency.
Live
Market Data
Monaco
Code Editor
Edge
Deployed
Next.js Python Groq Binance API Cloudflare Monaco
github.com/Quantis →
03
03
FastAPI · React · Playwright · Groq
DataReaper
Autonomous AI-powered privacy defense — it hunts your exposed data and forces deletion.
Active
DataReaper scans the web for your exposed personal data using advanced OSINT techniques and AI-powered identity resolution. It doesn't flag findings and stop there — it automates the full GDPR, CCPA, and DPDP legal workflows to force deletion from data brokers, people-search engines, and aggregators.
No lawyers, no manual opt-out forms. Playwright handles the browser automation, Groq drives the identity resolution, and PostgreSQL tracks every deletion status. It treats your privacy like a system invariant that must be maintained.
OSINT
Powered
GDPR
Automated
Auto
Deletion
FastAPI React Playwright Groq PostgreSQL
github.com/DataReaper →
Tech Stack
5 Categories
Languages
TypeScript
JavaScript
Python
C++
Frontend
React
Next.js
Tailwind CSS
HTML5
Backend
Node.js
Express
FastAPI
PostgreSQL
Redis
AI / ML
PyTorch
TensorFlow
scikit-learn
LangChain
RAG
Infra & Tools
Docker
Git
GitHub Actions
Cloudflare
Linux
GitHub Activity
Live
3
Flagship Projects
5+
Languages
10+
Frameworks Used
PRs Auto-Reviewed
GitHub Stats GitHub Streak Top Languages
Contribution Activity
Contribution Snake
Let's Connect
Let's build something
worth shipping.
Open to collaborations, side projects, and conversations about AI-powered systems that actually do something useful. If you have a hard problem and a bias for action, let's talk.
AI / ML Dev Infrastructure Privacy Tech Trading Systems Open Source
Reach out
harshsinghal1972@gmail.com linkedin.com/in/harshsinghal04 github.com/NA0XY
HARSH SINGHAL
GENAI · FULLSTACK · ML · 2025
BENGALURU, INDIA