Anup Singh

Implemented an automated Python script to streamline recurrent password changes across 50 servers, saving the team 3 hours per month. Reduced daily database check time by 50% through implementation of an automated process. Developed shell scripts that monitors, and reports Database related bottlenecks.

Led LLM Selection & User Engagement Strategy (Ideation Phase) Analyzed open leaderboards to select a large language model (LLM) capable of understanding user intent. Developed the chatbot interface and created 50 data points from scratch to fine-tune the LLM (Llama 7B) for Named Entity Recognition (NER). Increased NER accuracy from 20% to 80% through targeted data creation and fine-tuning. Developed a knowledge base that enabled the LLM to make 30% more accurate planning decisions compared to the base model, enhancing its overall effectiveness. Leveraged Google Colab for a serverless backend, reducing infrastructure costs during the Minimum Viable Product (MVP) phase by 100%.

Achieved 99.7% uptime for all Oracle Databases using proactive monitoring, recovery, and rman backup skills. Tested shell scripts for Switch over automation which led to 98.67% of improvement in Monthly DR Drills. Developed workflow to analyse the Active Session History of database and pinpoint the root cause of deadlocks in databases. Consistently exceeded client expectations, earning two Performer of the Month awards within a year.

Relevant Coursework: Dog Breed Classification, Online Retail Customer Segmentation, Sentiment analysis on Customer Reviews, Sales Forecasting on Walmart retail data.

Relevant Coursework: Notes Android App, Udaipur Travel guide, Ported Arrow OS for android phone on Cloud VM.