Joshua Suresh — AI Engineer

AI Engineer from Coimbatore, India specializing in Agentic AI Systems, LLM Orchestration, and large-scale information extraction. Currently an AI Engineering Intern at Infynd. B.Tech Artificial Intelligence & Data Science student at Karpagam Academy of Higher Education (2023–2027), CGPA 8.56.

Work Experience

AI Engineering Intern — Infynd (Feb 2026 – Present)

  • Built a two-stage company-people extraction pipeline processing 2.5M+ companies using GLiNER and Gemma4-E4B via vLLM, yielding 1.2M+ people records.
  • Built an agentic AI system using LangGraph to autonomously crawl company websites and extract structured company information.
  • Built data quality and QA workflows using SQL and Python to identify duplicates, missing records, and extraction inconsistencies across large-scale pipeline outputs.
  • Built a job title normalization system using vector embeddings and a vector database to map inconsistent, multilingual job titles to a standardized taxonomy.

AI Intern — Stepping Edge (Jun 2024 – Dec 2024)

  • Developed REST APIs for web scraping and structured product data extraction from e-commerce platforms, enabling automated catalog generation.
  • Built backend services to manage chatbot conversation history and multi-turn session continuation using unique session IDs.
  • Enhanced LLM chatbot performance using LangChain with few-shot prompting and improved query handling.
  • Researched and prototyped RAG, Knowledge Graph integration, and image similarity search for intelligent product discovery.

Projects

DomainIntel — Agentic Company Intelligence System

Designed a 9-node LangGraph pipeline to autonomously crawl, extract, and structure B2B company intelligence. Includes social URL extraction, redirect resolution, and a SaaS-style review dashboard. Technologies: Python, LangGraph, Crawl4AI, Qwen2.5, Pydantic v2. Winner — InFynd AIM 2025 Hackathon.

Satellite-Based Air Quality Downscaling

Developed ML models converting low-resolution satellite pollution data into high-resolution local air quality maps using Random Forest and CNNs, achieving over 90% validation accuracy. Technologies: Python, Random Forest, CNN, Satellite Data. Smart India Hackathon 2024 Finalist (ISRO Problem Statement).

Rural AI Assistant — Offline-First Voice AI

Offline-first AI assistant for rural shopkeepers and field workers. Voice-driven stock tracking and sales logging that runs entirely on-device with no internet required. Technologies: Python, Offline LLM, Voice Interfaces, SQLite.

Skills and Technologies

Python, FastAPI, SQL, MySQL, LangGraph, LangChain, vLLM, Ollama, GLiNER, spaCy, Crawl4AI, Asyncio, Docker, Git, Linux, TensorFlow, Keras, Scikit-learn, Pandas, NumPy, ChromaDB, Pinecone, HuggingFace, CrewAI, Microsoft Azure AI, IBM Data Science Professional.

Skills: LLM Applications, Agentic AI, AI Automation, Information Extraction, Web Scraping, ETL Pipelines, Backend API Development, SQL Analysis.

Achievements

  • Winner — InFynd AIM 2025 Hackathon (AI & B2B Data Intelligence Solutions)
  • Finalist — Smart India Hackathon 2024 (ISRO Problem Statement)
  • Finalist — Rakathon 2024 by Rakuten India (Top 100 out of 8,000+ teams)
  • Campus Ambassador — GeeksforGeeks 2025

Contact Joshua Suresh

Email: joshuasuresh08@gmail.com | GitHub: github.com/joshhuu | LinkedIn: linkedin.com/in/joshua08 | Portfolio: joshuas.tech | Location: Coimbatore, Tamil Nadu, India

$
josh.ai
22:33:34
~/josh $

Joshua S

// AI Systems · Full Stack · Ships Real Products

Obsessed with building AI systems that solve real problems. Deeply interested in Agentic Architectures

Available for freelanceOpen to Full-time Opportunities
Photo of Joshua S
Animated avatar of Joshua S waving
Click to flip