Harsha Vardhan

@harshavk07

Data Scientist

Índia
Inglês
Algumas informações são exibidas no idioma inglês.
Sobre mim
Hi! I’m a Data & AI professional with experience in Python, SQL, data analysis, machine learning, and data visualization. I help clients turn raw data into meaningful insights through clean, accurate, and well-structured solutions. I can assist with data cleaning, analysis, dashboards, Excel, SQL queries, Python projects, and ML tasks. I focus on quality, clear communication, and meeting deadlines. Whether you need help with a small data task or a complete project, I’m here to help. Let’s turn your data into actionable insights!... Saiba mais

Habilidades

h
harshavk07
Harsha Vardhan
offline • 

Conheça meus serviços

Desenvolvimento de chatbots de IA
I will build a rag pipeline and data engineering solution
A partir deUS$ 25 /projetoMais detalhes

Portfólio

Experiência profissional

YOGA_SANGEETA INC

GenAI / Data Engineering Intern

YOGA SANGEETA INC • Meio período

Jun 2026 - Aug 2026 • 2 mos

Built and optimized GenAI data pipelines supporting large-scale document processing and RAG applications. Designed a cache-first prompt architecture, task-based model routing, deterministic validation guards, automated ingestion, and machine-checkable quality controls. Processed 9,000+ source documents and produced 4,000+ RAG-ready Markdown documents. Reduced measured LLM inference cost by ~90% through call reduction, caching, routing, and deterministic validation while maintaining automated quality checks.

Publicis_Sapient

Data Science Intern

Publicis Sapient • Meio período

May 2025 - May 2025 • 0 mos

Developed a digital twin for a European FMCG client to improve demand forecasting, inventory optimization, and anomaly detection. Simulated multi-source data including sales, inventory, weather, and customer sentiment for machine learning training. Built XGBoost demand forecasting models using pricing, promotions, and sentiment as external drivers. Implemented EOQ and Deep Q-Network (DQN) approaches for inventory optimization, achieving over 95% service level while minimizing costs. Designed anomaly detection for demand, inventory turnover, forecast errors, service levels, and cost deviations. Built the solution using Python, Pandas, NumPy, Matplotlib, Scikit-learn, LLMs, Selenium, FastAPI, React, Docker/Podman, and Azure. Containerized and deployed the application dashboard to Azure for real-time accessibility.