Open to opportunities

Leena Harpal User

@leenaharpaluser

I build production RAG and machine learning systems that improve decisions, workflows, and operational knowledge access.

India

Message

What I'm looking for

I'm looking to build and ship production AI, GenAI, and RAG systems where I can solve meaningful business problems, lead technical delivery, and improve real-world workflows with reliable, measurable AI.

I've built production GenAI systems for steel plants, fashion-tech teams, and banking clients, turning complex operational and analytics workflows into practical AI products.

As an Independent AI/ML Engineer, I delivered a Plant Operations Knowledge Assistant and Maintenance Copilot that indexed 200+ SOPs, safety manuals, and maintenance logs, reducing lookup time from roughly 30 minutes to under two minutes. I also built an LLM-powered fashion recommender and a multi-agent trend researcher using LangChain, LangGraph, semantic search, Pydantic, and Langfuse tracing.

At Capgemini, I led a 16-engineer ML team delivering 15 predictive models for Standard Chartered Bank, improving forecast accuracy by 25%. I also deployed a GPT-3 SAS-to-Python converter adopted by DBS and ANZ, contributing to a 20% sales uplift.

Earlier, I built cyber-alert triage and anomaly-detection systems at JLL, and automated data preparation, chatbots, and dashboards at Amelia. I bring hands-on experience across Python, RAG, agentic workflows, AWS deployment, evaluation, and end-to-end ML delivery.

Experience

Work history, roles, and key accomplishments

Independent AI/ML Engineer & Ex-Founder

Self Employed

Jan 2023 - May 2026 (3 years 4 months)

Built RAG-based knowledge assistants and LLM-powered recommendation systems for manufacturing and retail clients. Founded a sustainable fashion startup, scaling it to INR 9 Lakhs annual revenue.

LangGraph Qdrant fastAPI Docker RAG LLM As Judge Pydantic Semantic Search

AI/ML Engineer - Team Lead

Capgemini

Feb 2022 - Dec 2022 (10 months)

Led a 16-member ML team to deliver 15 predictive models for Standard Chartered Bank, improving forecast accuracy by 25%. Designed a GPT-3 powered SAS-to-Python converter adopted by Fortune 500 banks, driving 20% sales uplift.

XGBoost Gradient Boosting Machine Learning GPT 3 NLP Flask Team Leadership

Machine Learning Engineer

Jones Lang Lasalle

Mar 2021 - Feb 2022 (11 months)

Built ML classifier to triage cyber alerts, reducing analyst workload by 30%. Developed real-time anomaly detection system using Deep Learning Auto Encoder, improving security response times by 35%.

Machine Learning Deep Learning Encoder Tableau Power Automate Cybersecurity

Data Analyst

Amelia

Jul 2019 - Mar 2021 (1 year 8 months)

Automated data preprocessing tasks, reducing manual effort by 30%. Built and optimized chatbot solutions using NLP techniques and designed data visualization dashboards using Python and Tableau.

SQL Python NLP Tableau Data Visualization Chatbot