Open to opportunities

Yash Sharma

@yashsharma20

Message

Machine learning engineer building automated, multi-country forecasting and LLM data pipelines.

Zimbabwe

Message

What I'm looking for

I want to build end-to-end ML systems that automate data pipelines and forecasting, plus work on LLM/NLP and computer vision models with strong optimization like quantization.

At TESCO, I built a time-series forecasting framework for the Czech Republic using SARIMA, Random Forest, and XGBoost, then automated holiday/payday uplift models, promo-mechanics features, and EDA/data preprocessing over 5+ years of sales data. I later scaled the same forecasting architecture to Hungary and Slovakia by automating ETL pipelines and integrating calendar mappings, exchange rates, and budget models.

Before that, at Turing, I helped develop Google DeepMind’s LLM models by creating complex datasets, building careful preprocessing workflows, and designing interactive multi-dimensional visualizations. I also worked on classification-focused modeling with DenseNet121 and ResNet, training DenseNet121 on Mini-ImageNet and applying pruning and quantization to cut model size from 27 MB to 8 MB with a small accuracy drop.

Experience

Work history, roles, and key accomplishments

Current

Machine Learning Engineer

Current

Tesco

Jun 2025 - Present (1 year 2 months)

Built a time-series forecasting framework for the Czech Republic using SARIMA, Random Forest, and XGBoost, improving accuracy by 12–18% across 30+ departments. Scaled the ML forecasting architecture to Hungary and Slovakia by automating ETL pipelines.

SARIMA Random Forest XGBoost ETL Time Series Python

Machine Learning Engineer

Turing

Sep 2024 - Mar 2025 (6 months)

Contributed to Google DeepMind's LLM models by creating datasets and preprocessing workflows. Trained DenseNet121 on Mini-ImageNet and applied quantization to reduce model size from 27 MB to 8 MB.

DenseNet121 ResNet Quantization Pruning LLM Python

Machine Learning Engineer

Regal Rexnord

Mar 2024 - Jun 2024 (3 months)

Developed a Chatbot with memory using RAG and LLM for PDF retrieval. Generated commodity codes using prompt engineering, NLP, and fuzzy matching. Optimized time forecasting models for price predictions.

RAG LLM NLP Fuzzy Matching ARIMA SARIMA Python