Akshay Patil

An enthusiastic Gen AI professional experienced in leveraging AI/ML and DL capabilities. Projects handled recently using AI, Machine learning and Deep Learning capabilities: 1. Deployed Machine learning models for Translation, Transliteration, Named Entity Recognition, Domain Classification, Sentiment Analysis, Speech Recognition and Speaker Diarization, Relevancy Score, Event Calender, Deepfake detection, Object Detection for client National Intelligence grid's project Sudarshan ( Dec 2022- Feb 2025) 2. Advanced RAG System for E-news Fact Checking & Summarization ( Mar 2025- ) Designed and deployed an on-premise advanced RAG system for structured system data, enabling fact-checking and summarization of E-news. Implemented guardrails and query keyword extraction with Elastic VectorDB (HNSW indexing), similarity search, and cross-encoder reranking, ensuring accurate, safe, and consistent LLM responses, optimizing performance and reliability in production. 3. Sentiment Analysis (Mar 2022-Oct 2022) Implemented supervised classification algorithms to analyze sentiments of tweets. Performed NLP preprocessing including stopword removal, tokenization, stemming, and lemmatization. Evaluated models using F1-score to ensure balanced performance across sentiment classes. Integrated Tweepy API for data collection and used AWS for scalable processing and storage. Work activities of latest projects: 1. Understanding Business scope, problem statement and data gathering from client. 2. Data cleaning, preparation and visualization activities. 3. Steady state data extraction from raw sensor data. 4. Creating Pipelines for Data Preprocessing activities. 5. Feature Engineering, Feature selection, Machine learning and Deep learning Model building activities. 6. Evaluating and improving model performance by fine tuning hyperparameters. 7. Creating Pipelines for Model buildi