CS Engineering student at SRM University Delhi-NCR (AI & Data Science), currently interning at Beagle Security as a Business Development Intern where I designed and built Sonar — an internal Apollo-alternative for extracting live LinkedIn data and generating ICP-qualified prospect lists at scale.
I build full-stack AI products end-to-end: from database schema to deployed UI, enrichment pipelines to LLM orchestration. My work is production-grade, observable, and fast by default.
Open To: SDE internships · AI/ML engineering roles · Full-stack product roles · Research collaborations
Languages
Frontend
Backend & Databases
Cloud, DevOps & Tooling
| Category | Tools |
|---|---|
| AI / LLMs | Groq, Gemini AI, OpenRouter |
| Scraping | Playwright, Firecrawl |
| Enrichment APIs | People Data Labs, Hunter.io, Apollo.io |
| Extensions | Chrome MV3 |
| Observability | PostHog |
Remote · Apr 2026 – Present
Designed and built Sonar — an internal Apollo-like lead intelligence platform for extracting live LinkedIn data and generating ICP-qualified prospect lists at scale.
- Architected a FastAPI backend on Vercel with Supabase PostgreSQL, handling auth, lead storage, and enrichment orchestration
- Built a Chrome Extension (MV3) for one-click LinkedIn profile and company page extraction with real-time backend sync via authenticated REST API
- Engineered a multi-stage email enrichment pipeline — website scraping → pattern inference → Hunter.io + Apollo API, processing hundreds of companies concurrently
- Integrated People Data Labs API for large-scale prospect discovery and bulk lead enrichment
- Built AI-powered workflows using Gemini AI, Groq, and OpenRouter for ICP scoring, business model classification, and company intelligence analysis
- Delivered a React dashboard with 10+ modules — Company Intelligence, Prospect Search, ICP Scoring, Email Sequences, Analytics, Administration
- Implemented PostHog analytics for product usage tracking, funnel analysis, and feature adoption
27 automation scripts + 2 terminal menus + 18 scheduled LaunchAgents — built to reclaim space on a 256GB Mac and make a 1TB external SSD do all the heavy lifting.
git clone https://github.com/sxrxvxnn/mac-ssd-toolkit
cd mac-ssd-toolkit && ./install.sh| Attribute | Detail |
|---|---|
| Stack | zsh, macOS launchd, Homebrew |
| Result | 13GB+ freed on internal drive, nightly automated backups |
| Article | How I freed 15GB on my 256GB MacBook |
| Repository | github.com/sxrxvxnn/mac-ssd-toolkit |
Sonar — AI-Powered B2B Lead Intelligence Platform
Internal Apollo-alternative built for Beagle Security. Autonomously discovers, classifies, enriches, and validates B2B prospects at scale using multi-source intelligence pipelines and LLM-based decision making.
| Attribute | Detail |
|---|---|
| Stack | React, TypeScript, Vite, FastAPI, Supabase PostgreSQL, Groq, Gemini AI, Firecrawl, PDL, Playwright |
| Scale | Multi-source enrichment across PDL, Hunter.io, Apollo.io with deduplication and entity resolution |
| Performance | Async enrichment workers, streamed LLM outputs, <2s lookalike scoring |
| Security | Row-level security via Supabase, JWT auth, API key rotation |
| Repository | github.com/sxrxvxnn/leadgen-platform |
Lookalike Search Engine — Semantic B2B Company Matching
LLM-guided semantic similarity scoring — single structured Groq call ranks companies against a target by industry, business model, and tech signals. <2s end-to-end.
| Attribute | Detail |
|---|---|
| Stack | TypeScript, Groq, Supabase, PostgreSQL |
| Performance | Single Groq API call with structured JSON output; <2s end-to-end |
| Repository | github.com/sxrxvxnn/leadgen-platform |
LinkedIn DM Verification Pipeline — Playwright Scraping Worker
Playwright pipeline verifying decision-maker profiles on LinkedIn — handles SPA rendering, hash-based class names, and cross-validates against company website records.
| Attribute | Detail |
|---|---|
| Stack | Python, Playwright, FastAPI, PostgreSQL |
| Scale | Async batch processing, designed for queue-driven horizontal scaling |
| Repository | github.com/sxrxvxnn/leadgen-platform |
Sonar Chrome Extension — LinkedIn Data Extractor (MV3)
One-click LinkedIn profile and company page extraction into Sonar, with real-time backend sync via authenticated REST API.
| Attribute | Detail |
|---|---|
| Stack | Chrome Extensions Manifest V3, JavaScript |
| Repository | github.com/sxrxvxnn/leadgen-platform |
IBM
Databricks
Udemy
- 🌱 Created
masterin sxrxvxnn/sonar-intelligence on 2026-07-29 - ⭐ Starred sxrxvxnn/mac-ssd-toolkit on 2026-07-28
- ⭐ Starred sxrxvxnn/claude-plugins on 2026-07-28
- ❗ Opened issue claude-plugins: complementary hooks for security, cost tracking, and workflow in JuliusBrussee/caveman on 2026-07-28
- 🌱 Created
add-claude-pluginsin sxrxvxnn/awesome-claude-code on 2026-07-28 - 🍴 Forked hesreallyhim/awesome-claude-code on 2026-07-28
- 🌱 Created
mainin sxrxvxnn/claude-plugins on 2026-07-28 - 🌱 Created
dependabot/npm_and_yarn/dashboard/motion-12.42.0in sxrxvxnn/leadgen-platform on 2026-07-28 - 🌱 Created
dependabot/npm_and_yarn/dashboard/posthog-js-1.396.0in sxrxvxnn/leadgen-platform on 2026-07-28 - 🌱 Created
dependabot/npm_and_yarn/dashboard/axios-1.18.1in sxrxvxnn/leadgen-platform on 2026-07-28
Setup: Install WakaTime VS Code extension → add
WAKATIME_API_KEYsecret → stats appear automatically
current_focus:
role: Business Development Intern @ Beagle Security
building:
- Sonar — AI-powered B2B lead intelligence platform
- LinkedIn DM verification pipeline (Playwright + Python)
- Lookalike search engine with structured LLM scoring
learning:
- Advanced LLM agent architectures
- Distributed enrichment pipeline engineering
- Production observability for AI systems
open_to:
- SDE internships at product-led companies
- AI/ML engineering roles
- Full-stack product engineering positions
"The best engineers don't just write code — they build systems that make the impossible routine."