Harpreet Kour

About Harpreet Kour

I am currently working as an AI/ML Engineer (Trainee) at Hackerkernel, where I am gaining hands-on experience in building AI-driven systems and working with real-world data problems.I am particularly interested in Artificial Intelligence, Machine Learning, and Data Science, and I enjoy building intelligent solutions that automate workflows and generate insights from data. At Hackerkernel, I am currently working on an AI-based automated follow-up system, focusing on designing logic, integrating AI capabilities, and improving automation processes.My key areas of interest include:• Machine Learning• Artificial Intelligence & LLMs• Python for Data Analysis• Data Processing & Automation• Building AI-powered applicationsI am continuously improving my skills in Machine Learning, Data Structures & Algorithms, and AI system design, while exploring how AI can be applied to solve real-world problems.I am always open to connecting with professionals in AI, Machine Learning, and Data Science to learn, collaborate, and grow in this field.

Experience

1. Ai Ml Engineer

   HackerKernel

   Dec 2025 — Present · Bhopal, IN

   I build AI-powered systems that think, learn, and act — not just tools that respond.I engineered a full-stack intelligent outreach automation platform from the ground up — a system that handles end-to-end B2B LinkedIn sales conversations autonomously, with the quality and nuance of a seasoned CEO.What I built:A two-pipeline RAG architecture with real-time intent classification, dynamic state detection, and portfolio-aware response generation — deployed via Docker, integrated with a Chrome extension working across LinkedIn and AimFox.Key highlights:→ Dual ChromaDB RAG pipeline with semantic search, metadata filtering, and sliding window context management at scale→ Intent & state detection engine combining RAG classification with LLM fallback across positive, negative, objection, and neutral flows→ Portfolio retrieval system with tech-aware filtering and hallucination prevention — AI never invents links or shows irrelevant work→ Feedback loop where approved/edited responses re-embed into ChromaDB — only quality-controlled examples influence future AI behavior→ Zoho Bigin CRM integration, RAG Management System with bulk upload and vector lifecycle management→ Chrome extension supporting both LinkedIn and AimFox with platform-aware DOM scrapingStack: Python · FastAPI · PostgreSQL · ChromaDB · OpenAI GPT-4o · Docker · JavaScript · Zoho Bigin API · Vector EmbeddingsI don\'t just use AI — I build the infrastructure that makes AI reliable, controllable, and production-ready.

Education

• Hemchand Yadav Vishwavidyalaya, Durg

  Bachelor of Science - BS, Microbiology, General

  2022 — 2025