Jaymin Joshi
Development
Gujarat, India
Skills
Machine Learning (ML)
About
Jaymin Joshi's skills align with Programmers (Information and Communication Technology). Jaymin Joshi appears to be a low-to-mid level candidate, with 2 years of experience.
Work Experience
Machine Learning Engineer
SoftmaxAI
April 2022 - Present
- India • Develop and maintain AI Backend on a content management website having thousands of active users. Developed automated pielines to generate summaries, titles, descriptions, chapters, etc. for videos and documents (PDF, docs, sitemaps, etc.) . Developed chatbot to interact with the user content using RAG pipeline powered by LLMs. • Developed several chatbots using RAG pipelines using Qdrant, Pinecone, Chroma, etc. vector databases and LLMs such as GPT, Gemini, Llama, Mistral, etc. that answers using custom knowledge base and has streaming response support. • Developed a chatbot that replicated a technical support professional for users having techincal issues with a platform. It analyses the user provided error logs and provide response based on large dataset of similar historical conversations using advanced RAG techniques. The chatbot endpoint was deployed entirely using Google Cloud. GCP services used were VertexAI, Vector Search, Text Bison LLM, Cloud Run. • As per client requirements, developed a desktop app using Tkinter that analyzes Financial documents uploaded by user and generate reports and answers specific questions similar to a seasoned credit analyst. This software was built using Advanced RAG techniques and prompt engineering with continuous monitoring of responses and iterative improvements. • Optimized AI Backend of a mobile retail app that uses Object Detection, Image Classification, Optical Character Recognition (OCR) with YOLOv5, FastAI, PaddleOCR respectively to calculate various metrics for compliance. • Developed fully automated data pre-processing and model training pipeline for object detection. The pipeline was fully serverless and deployed on AWS. The cloud infrastructure was deployed and managed using Terraform. AWS architecture included Lambda, S3, Sagemaker. • Deployed forecasting pipeline on GCP (VertexAI, Cloud Functions) • Developed Recommendation Engine on AWS Personalize for a mental health based mobile app having thousands of active users. • Developed and deployed optimised object detection pipeline on the edge with NVIDIA Deepstream and Jetson Nano. ? Certifications AWS Certified Solutions Architect - Professional Demonstrating mastery in designing and deploying scalable, resilient, and secure AWS solutions. Expertise in advanced architectural principles, multi-tier application design, and cost optimization. Trusted to architect complex cloud-native applications meeting business objectives with efficiency and innovation. ? Publications • Experimental investigation of different NN approaches for tool wear prediction based on vision system in turning of AISI 1045 steel, International Journal on Interactive Design and Manufacturing (IJIDeM)
Education
Nirma University
Bachelor of Technology
January 2018 - January 2022
