Complex problem-solver with analytical and driven mindset. Dedicated to achieving demanding development objectives according to tight schedules while producing impeccable code.
Overview
8
8
years of professional experience
Work History
Senior AI Engineer
Konverge.AI Pune
04.2024 - Current
Conceptualized and implemented an automated pipeline to iteratively generate data points to improve the accuracy of the fine-tuned LLM model for the question-answering task, reducing 50% of manual retraining.
Researched on the Invoice-processing workflow to inject LLM Agents in the loop for automating the tasks, making the workflow more efficient
Created different LLM agents and tools and integrated with the Emergence Orchestrator platform.
AI Engineer
04.2023 - 04.2024
Conceptualized a state of the art Search and Question Answering platform, leveraging LLMs and advanced RAG techniques, with 90% accuracy by driving a team of five; the platform has attracted new clients for the company
Created an AI Avatar web-application using LLMs and GAN techniques for a not-for-profit organization to assist job-seekers prepare for interviews, thus allowing for a vast geographical reach – Lead a team of five to develop a RAG based Question-answering web-application for any kind of documents (PDFs,CSVs, Databases etc.)
Software Engineer
02.2022 - Current
Initiated a natural language search engine for an AI start-up in Governance with a 92% accuracy that made it more customer friendly for the client and proved to be critical for their future revenues.
Developed an AI-based coaching and evaluation platform by leveraging advanced OCR and LLMs to empower aspiring entrepreneurs by providing human-like mentoring.
Developed a product to showcase the product capabilities of Konverge.AI that raised client engagement by 50%.
Assistant Manager
Jio Platforms Limited Remote
08.2023 - 10.2023
Founding member of AI-ML team at Jio and worked with cross-functional teams to create the processes for product development and liaised as the PoC (Point of Contact) between the outsourced company and our AI/ML team.
Developed an automated computer vision model for housekeeping compliance that reduced supervisory activity by 80% and was later approved for marketing.
Created the customer analytics models for restaurants to effectively manage the customers real-time, which was successfully tested, demonstrated to management and approved for marketing.
Trained object detection(YOLO) and tracking(Deep SORT) model to count passengers boarding a bus, helping managing the bus occupancy limit real-time.
The end-to-end pipeline was deployed on both cloud and edge device – Achieved accuracy of 0.92% in the anomaly detection of working appliances in smart home analytics
COEP Technological University
08.2016 - 05.2020
Implemented a deep learning model for Image Captioning using Convolutional Neural Networks .The model was trained on the COCO dataset.
Verified the relation of Fins with heat transfer by data driven Neural network methods. This project uses neural networks to confirm established relationship of heat transfer between Fins
The model was trained on a small neural network on the data consisting of various parameters of the heat transfer.
Performed tool Condition monitoring of Multi-Point Cutting Point Tool using Machine Learning Approach.
Transformer
Orphanage Home
08.2023 - Current
Projects – RAG Apps: Created a number of apps leveraging Generative AI and advanced RAG techniques
TechStack:LangChain, LlamaIndex, AWS-Bedrock, HuggingFace, LangSmith, LangServe,OLLama, vLLM, LangGraph etc
Fine Tuning of Open-source LLMs: Fine tuned a number of open-sourced LLMs using QLoRA and LoRA PEFT techniques. Also, aligned the models using SFT, ORPO, DPO techniques
LLMs: Mistral,Llama,Phi-3,GPT-Neo etc
Created end-to-end deep learning model pipeline from training to model deployment
From the scratch implementation of NLP & CV GANs,DeepSort,LLama,VAE,CNN, YOLO,Stable-Diffusion,RLHF etc
Achievements and Experiences – Contributed for the research project titled - “To review fraudulent reviews provided by the patients using GPT” – Actively contributes to OSS - DSPy, vLLM, Streamlit, etc
And various other ML/DL based libraries
Member of the organizing committee of the Founding Day Programme of Konverge.AI : (April 2023) – Konverge.AI - Initiated and led a team to commence CSR activity that focuses on providing basic needs and study materials to an, ) – Member Robotics Study Circle,COEP - Head of the Mechanical team and was part of the team that developed the All-Terrain Robot
Education
BTech - Mechanical Engineering
College of Engineering
Pune, MH
05.2020
Minor in Computer Science
College of Engineering
Pune, MH
05.2020
Skills
Skills Summary
Languages:
Python, C, C
Full Stack Development: FastAPI,Streamlit,Flask,SQLALchemy, MySQL, MongoDB, Redis,RabbitMQ, PostgresQL,Multithreading