Aug 2024 – Dec 2024 · 4 mos · Bengaluru, Karnataka, India · On-site

• Extracted product level filters from its Image + Description across 1000 categories, using OpenAI GPT-4o model with Vision capability
• Created an highly extensible MLOps framework for ML model inference using TF-serving. Reduced latency from ~400ms to ~5ms per request
• Build Low-code MLOps platform with MLFlow to enhance the Data-science lifecycle (deployment + inference).
• Created Databricks registry and ensured 100% model artifacts & metadata registration for version tracking by DS team

Jan 2024 – Oct 2024 · 9 mos

• Working on LLM products and building highly scalable ML Infrastructure

Aug 2022 – Jan 2024 · 1 yr 5 mos

• ML-INFRASTRUCTURE:
• Enhanced security of 4000+ deployments by using Identity and Access management (IAM) with minimal downtime. Achieved using bulk update & batched migration.
• Created Internal filesystem to provide interactive APIs across AWS, Azure, GCP & Local system. Currently utilized in more than 90% of live deployments
• Migrated 300+ unsecured docker images to secure docker, added low privileged user, removed critical vulnerabilities. Published a set of secure base dockerfiles which can be inherited as per need
• Reducing the cost of deploying in-house LLMs by adding optimizations using Continuous and Dynamic batching. (in-progress)
• DATA-SCIENCE:
• Implemented AutoML for text similarity, which achieved 100% automation in new client onboarding by adding support for multiple sentence encoders including MPNet, MUSE, Distilbert in a generic framework
• Leveraging Prompting and OpenAI GPT-4 to generate Playbooks for agents to produce next responses given a Case conversation and User query.
• Created Quality Management product which detects customer care Agent’s performance with a revenue growth of 100% every quarter. Trained wide range of models including NER, Multilabel Classification, FAISS similarity

Aug 2020 – Jan 2024 · 3 yrs 5 mos

• Training or Fine-tuning State-of-the-art Natural Language Processing models.
• Deploying trained ML model on Kubernetes cluster.
• Ensuring Reliability, Scalability, and Availability of the deployed models.
• Technologies:
• Python, Keras, Tensorflow, Pytorch, Docker, Kubernetes

Jun 2020 – Jul 2020 · 1 mo · India

• Full-stack development of the entire website and periodic deployment on the server.
• Worked on features like Online chats, Video calling, and Live Counselling sessions.
• Developed a module that recommends possible Career/College(s) to Students.
• Technologies:
• Codeigniter, Bootstrap, PHP, HTML/CSS, Javascript, MySQL

May 2019 – Jul 2019 · 2 mos · Greater Hyderabad Area

• Worked on improving the Accessibility of a Web Application.
• Resolved the accessibility issues which improved compliance for visually impaired people.
• Designed the major landing pages from scratch.
• Technologies:
• HTML/CSS, Typescript, Angular, Microsoft SQL Server