Nilakshi Naphade

GenAI & NLP: Built LLM-based diagnostic tools (Streamlit, RAG) and automated translation/classification pipelines using HuggingFace & VertexAI. MLOps & Engineering: Defined unified Azure DevOps processes (CI/CD, Monitoring) and engineered robust PySpark/Databricks pipelines. Analytics: Developed predictive models (Residual Value), anomaly detection, and PowerBI dashboards for cost optimization. Strategy: Led vendor POCs for data lake solutions adopted company-wide.

- Data analysis using R scripts - Predictive maintenance and analysis using unsupervised machine learning algorithms viz. K-means - Anomaly detection for the real-time machines - Image detection using CNN based YOLO algorithm - Coursera online certification for deep neural networks (NN) and convolutional NN (CNN). - Hands-on experience with hyperparameterisation, and optimization for deep neural networks

Development and maintenance of web application based on React-JS and Python based Pyramid framework

Research assistant for the project "Vandalism Detection in Wikidata". Developing new approaches for detecting vandalized edits in Wikidata knowledge base.

Worked as a Software Engineer for the "Mediamorph" project team for 2 years. Development and maintenance in Java and PostgreSQL

Master Thesis (Knowledge bases and Data Science field)​: Topic: Vandalism Detection in Wikidata's Type Relations. A research based application for detecting vandalized edits in Wikidata's type statements. Designed and developed machine learning based algorithm using Python and Java.

"American Sign Language Recognition", a research based project for physically challenged people. Technologies: MATLAB, mySQL. Published a review paper titled 'Recent Developments in Sign Language Recognition: A Review' International Journal (IJACECT).