Open to opportunities

Abhishek Dureja

@abhishekdureja

Message

Data Scientist building agentic AI systems for automation, forecasting, and recommendations.

India

Message

What I'm looking for

I want to build production-ready agentic AI systems that automate business workflows, improve forecasting accuracy, and turn data into actionable decisions—using strong ML practices, clean APIs, and reliable deployment.

I’m a Data Scientist specializing in Agentic AI, machine learning, and predictive analytics. I’ve built and deployed 10+ AI-powered workflows that automate reporting, campaign analysis, and recommendations using LangGraph, OpenAI, and Claude, and I work end-to-end from model development to production deployment.

In my role, I also built a centralized AI-powered marketing intelligence platform integrating multi-channel advertising data, and I reduced manual reporting and campaign analysis effort by ~50%. I develop forecasting models with Prophet and XGBoost (R score 0.88) and engineered recommendation systems using campaign performance and seasonal signals to drive actionable optimization.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Lyxel Labs Pvt. Ltd.

Apr 2025 - Present (1 year 3 months)

Built and deployed 10+ AI-powered workflows using LangGraph, OpenAI, and Claude to automate reporting, campaign analysis, and recommendations. Developed forecasting models with Prophet and XGBoost and engineered recommendation strategies for advertising optimization.

LangGraph OpenAI Claude AI Workflow Automation Marketing Intelligence Forecasting Prophet XGBoost Recommendation Systems Campaign Optimization

AI Intern

Affine Analytics

Dec 2024 - Mar 2025 (3 months)

Developed a real-time, customizable live streaming avatar using RAG for interactive streaming applications. Researched LLM quantization and designed Light RAG to improve retrieval speed and output precision.

RAG LLM Quantization Information Retrieval

AI Trainee

Intel

Jun 2023 - Jul 2023 (1 month)

Worked on an AI-driven air quality prediction system using Decision Trees on environmental sensor data, achieving 96.8% accuracy. Benchmarked predictive performance using Logistic Regression, KNN, Naive Bayes, and SVM.