|
With 6+ years of experience in ArtificaI Intelligence and Machine Learning, I specialize in
designing scalable, data-driven solutions
that deliver measurable business impact. My expertise spans Generative AI, Advanced Natural Language
Processing(NLP),
Recommendation Systems, Speech/Audio Processing and Computer Vision(CV), with a strong foundation in
building
scalable MLOps/LLMOps pipelines.
|
|
|
|
|
April 2021 - August 2023
Worked with Place Creation team at HERE Technologies to expand the coverage of HERE Maps by
17% across 9 countires by extracting geospatial data from websites using state of the art
machine learning, deep learning and Large Language Models generating over 10 million
high-quality records.
Crafted a scalable MLOps/LLMOps pipeline on AWS, leveraging
optimized CPU/GPU instance selection and model compression techniques, including Knowledge
Distillation and ONNX.
|
|
August 2017 - April 2021 Worked with the AI@Scale team at Fractal on building the platform solutions in machine learning, deep learning and big data. Worked with Fortune 500 clients such as Colgate , Autodesk and Procter & Gamble on projects involving forecasting, predictive analytics, big data technologies, speech and natual language processing using machine learning. Tools and Technologies : python, pytorch, tensorflow, pytorch lightning, docker, git, Fast API, django, Microsoft Azure, AWS, Airflow, Jenkins, Numpy, Pandas, Scipy, Scikit-learn, Statsmodel, Matplotlib, Jupyter, langchain, sentence transformers, huggingface |
|
|
|
|
May 2024 - August 2024
Worked with the Place Creation team to accelerate data preparation, training, and testing
for multiple machine learning and deep learning models by 60%
through streamlining routine processes and leveraging Large Language Models (LLMs) such as
Llama3 and OpenChat with carefully crafted prompts for feature extraction and model
evaluation. I also built an AI Chatbot to assist with onboarding and information retreival.
|
|
|
|
Developed pipeline for personalized text-to-image generation and editing using latent diffusion models (generative AI), incorporating LoRA finetuning for image generation and cross-attention guidance for image editing (pytorch / hugging face diffusers /stable diffusion) Tools and Technologies : python, pytorch, hugging face diffusers, stable diffusion, OpenCV, transformers |
|
Code Designed Large Language Model chatbot to assist students, faculty, and prospective applicants by providing detailed information about various aspects of university life and administration at Stevens Institute of Technology. Built using Langchain, Streamlit, and GPT-4, this chatbot offers a user-friendly interface and comprehensive details across a range of topics. Tools and Technologies : python, pytorch, hugging face, Streamlit, Ollama, Langchain, transformers |
|
Code Fine-tuned Llama2 chatbot using QLoRA, tailored to provide detailed information and recommendations about movies. The model is fine-tuned on the IMDB dataset, enabling it to generate informed and contextually relevant responses. Tools and Technologies : python, pytorch, huggingface, transformers |
|
Code Modeled a collaborative filtering-based recommender systems for personalized Instacart recommendations, comparing performance with TF-IDF, Singular Value Decomposition(SVD), and Bayesian Personalized Ranking(BPR) methods Tools and Technologies : python, pandas, numpy, matplotlib, lightFM, implicit |
|
Code Project involves forecasting total sales for a brewery using a dataset from Kaggle. The models used for forecasting include Linear Regression, Random Forest, and Decision Tree on PySpark. The project has been executed and tested on Google Cloud Platform's DataProc service. Tools and Technologies : python, numpy, matplotlib, pyspark, matplotlib, Google Cloud Platform |
|
Code This project is a restaurant chatbot designed using Langchain and GPT, with a frontend built in Streamlit and APIs managed by FastAPI. The chatbot interacts with customers to provide personalized food recommendations and resolve customer complaints. All data is stored in a SQLite database. Tools and Technologies : python, pytorch, hugging face, Streamlit, Ollama, Langchain |
|
|