class PrashantSuthar:
role = "AI/LLM Backend Engineer"
experience = "8+ years"
location = "India 🇮🇳"
def superpowers(self):
return {
"ai": ["LLM apps", "RAG", "AI agents", "structured outputs", "model evals"],
"backend": ["FastAPI", "Django", "async workers", "high-volume APIs"],
"extraction": ["100+ sites scraped", "anti-bot bypass", "document parsing"],
"automation": ["end-to-end pipelines", "browser automation", "scheduled workflows"],
"cloud": ["AWS", "Docker", "CI/CD", "serverless"],
}
def mission(self):
return "Convert raw, scattered, unstructured chaos → clean, reliable, business-ready data"I build intelligent backend systems that transform websites, documents, APIs, databases, and unstructured content into clean, structured, actionable data — with AI at the core of the pipeline, not bolted on as an afterthought.
|
LLM-powered systems with structured outputs, semantic search, RAG, intelligent agents, and human-in-the-loop feedback workflows. Pipelines that extract, clean, validate, and monitor data from websites, marketplaces, directories, documents, APIs, and databases. Multilingual model evaluation, response benchmarking, correction workflows, and quality-measurement systems. |
Scalable APIs, async workers, auth systems, queues, and integrations — engineered in Python for real production load. Automating repetitive data-collection, processing, validation, monitoring, and delivery — so humans never do robot work. Dockerized, monitored, CI/CD-driven deployments on AWS — built to run unattended and recover on their own. |
AI / LLM Engineering
Data Extraction & Processing
| Project | What It Does |
|---|---|
| 🤖 AI-Powered Data Intelligence Platform | Connects to websites, documents, APIs & databases → extracts, validates, and transforms information into structured business data |
| 📊 LLM Benchmarking & Review System | Human-in-the-loop platform comparing AI model responses, collecting corrections, and computing quality metrics |
| 🎙️ Multilingual Speech Data Pipeline | Scalable audio segmentation, transcription, validation, and cloud-based dataset delivery |
| 💰 Automated Price Monitoring | Collects product prices, matches products, detects changes, and produces pricing intelligence |
| 🌐 Large-Scale Web Extraction | Extraction workflows for 100+ websites — ecommerce, property portals, travel sites, directories, public data |
| ⚡ Business Workflow Automation | Multi-step browser + backend + data-processing automation that eliminates manual work |
|
Give your LLM a front desk. An MCP server that turns Claude Desktop, Claude Code, or any MCP client into an AI receptionist for a medical clinic — booking & rescheduling appointments, patient registration, pre-visit intake, reminders, with destructive actions held back for confirmation and clinical questions routed away from the model. |
🌾 AgroSageMulti-agent crop advisory for smallholder farmers. Ask a free-form question — "Should I irrigate my cotton this week?" — and an LLM orchestrator delegates to specialist tools (live weather, mandi prices, an agronomy rule engine), returning a short actionable advisory in the farmer's own language. Also runs fully offline with no LLM and no API keys. |
Installable skills that teach coding agents to scrape like an expert. Self-contained instruction packages for Claude Code, Cursor, Copilot & 20+ agents — covering the full life of a scraping project: site feasibility audits, selector-free structured extraction, broken-scraper diagnosis, and proxy cost analysis. One command to install: |
More projects: Flask REST Boilerplate · Booking Website Scraper · YouTube Data Scraper · COVID-19 Data Scraper
I'm open to collaborating on:
🤖 AI & LLM-powered applications · ⚙️ Backend systems & APIs · 🕷️ Web scraping & data extraction · 📄 Document-processing automation · 📊 Data monitoring & intelligence platforms · ☁️ Cloud infrastructure & workflow automation