Hitesh Patil

Actively looking for full-time positions as an AI Engineer/ Machine Learning Engineer/ Data Scientist/ NLP Engineer/ Data Engineer/ Software Engineer.

Academic Background: I graduated with a master's degree in Artificial Intelligence from The State University of New York at Buffalo. My academic journey has been enriched with a broad spectrum of subjects, including Natural Language Processing (NLP), Machine Learning, Deep Learning, and Algorithms Design. This comprehensive exposure has equipped me with a solid understanding and practical skills in AI and its applications. My Bachelor's degree in Computer Science from Ramrao Adik Institue of Technology, provided a robust foundation, sparking my passion for cutting-edge technologies and complex problem-solving in the realms of Artificial Intelligence and Machine Learning.

Technical Skills: My skill set is versatile and robust, encompassing proficiency in various programming languages such as Python, Java, C++, PLSQL, and SQL. I also possess expertise in a plethora of frameworks and tools, including PyTorch, TensorFlow, Keras, Langchain, LlamaIndex, VectorDBs(ChromaDB, Faiss, elasticsearch), MLFlow, PySpark, Pandas, Flask, MongoDB, and SpringBoot. Furthermore, I have hands-on experience working with cloud platforms like Oracle, Azure, Google Cloud Platform (GCP), Amazon Web Services (AWS), Atlas, OpenAI.

Professional Roles: My journey in the fields of NLP, generative AI, Large Language Models (LLMs), and machine learning has been shaped by a series of impactful roles:

Machine Learning Engineer at CIPIO.ai: At CIPIO.ai, I enhanced the content retrieval library by developing a system that uses cosine similarity for efficient retrieval, transforming user content into OpenAI CLIP embeddings stored on Qdrant clusters, and improving retrieval with metadata like object, audio, and text embeddings.

Data Science Intern at Von Roll: At Von Roll, I designed an advanced Retrieval Augmented Generation (RAG) chatbot for knowledge sharing across six teams, integrated LlamaIndex ReAct Agents for automating insights from SAP reports, and optimized report generation, reducing time by 60%, while also applying LSTM and GRUs for a 20% increase in forecast accuracy.

Software Development Engineer at AJIO.com: At AJIO.com, I developed Java-Spring Boot micro-services with Apache Spark for data extraction, created GST-based tax algorithms for 90% accuracy, deployed scalable micro-services on Kubernetes, and developed automation suites for "JioMeet" with CI/CD pipelines on Azure DevOps.

Web Developer at ThingsGoSocial: At ThingsGoSocial, I upgraded the existing system to a MERN stack-based web application, contributed to API development, built a single-page application using Redux, Webpack, and Babel, and developed micro-frontends to reduce server load.

Summary: Experienced AI professional with a strong passion for leveraging advanced technologies such as NLP, Machine Learning, and Large Language Models (LLMs) to tackle real-world challenges. My solid academic background in Artificial Intelligence is complemented by hands-on experience in diverse roles, including developing content retrieval systems, engineering advanced Retrieval Augmented Generation (RAG) chatbots, and crafting scalable micro-services. With a track record of enhancing system efficiency and accuracy, I am eager to contribute to innovative advancements in AI and software development, driving impactful solutions in the industry.

I am highly enthusiastic about exploring opportunities in this space. If you have relevant openings or are interested in having a conversation on these topics. Feel free to check out my Resume and drop me an e-mail. I'd be happy to connect for a quick call to discuss!

~ Email | Resume | Github | LinkedIn ~

Feb '24	Joined CIPIO Inc. as a Machine Learning Engineer
Jan '24	Graduated with my Masters in AI from UB
Aug '23	Started my final semester(Fall 2023) at UB.
May '23	Joined Von Roll USA Inc. as a Data Science Intern
Jan '23	Started my second semester(Spring 2023) at UB.
Aug '22	Started my graduate studies at State University of New York at Buffalo(UB) - Fall 2022 Semester.
Nov '20	Joined AJIO as a Software Development Engineer
Oct '20	Completed my Bachelor of Engineering(BE/Btech) in Computer Engineering from RAIT.
Aug '20	Joined Things Go Social as a Web Developer Intern

State University of New York at Buffalo

Master of Science | Artifical Intelligence
Aug '22 - Dec '23
Coursework:
• CSE-574: Intro Machine Learning
• CSE-555: Intro Pattern Recognition
• CSE-635: Natural Language Processing and Text Mining
• CSE-531: Algorithms Analysis and Design
• CSE-546: Reinforcement Learning
• EE-559: Big Data Analytics
• CSE-540: Machine Learning & Society
• EAS-595: Fundamentals of AI
• EAS-501: Introduction to Numerical Mathematics for Data Scientists
• CSE-568: Robotics Algorithms

Ramrao Adik Institue of Technology, Navi Mumbai

Bachelor of Engineering | Computer Engineering
Specialization: Artifical Intelligence
July '16 - Oct '20

	Machine Learning Engineer \| CIPIO Inc. McLean, VA, USA Feb '24 - Present NLP Search based Content Retrieval System • Contributed to CIPIOs content retrieval library to procure related content stored on AWS S3 bucket, optimizing data access and retrieval processes. • Designed and implemented a retrieval system leveraging cosine similarity search between user-entered queries and content embeddings to ensure precise and relevant content matching. • Meticulously scripted the conversion of user content into OpenAI CLIP embeddings, storing these embeddings on Qdrant clusters to facilitate efficient retrieval. • Enhanced retrieval efficiency by over 80% by incorporating additional metadata, including object embeddings per keyframe using Faster R-CNN, audio embeddings for video content using OpenAI Whisper, and text embeddings for captions with OpenAI CLIP.
	Data Science Intern \| Von Roll USA Inc. Schenectady, NY, USA May '23 - Aug '23 Internal In-House Chatbot for Knowledge Sharing • Engineered an advanced Retrieval Augmented Generation (RAG) chatbot to facilitate knowledge sharing across six internal teams, improving information accessibility and collaboration. • Integrated LlamaIndex ReAct Agents to automate the extraction and analysis of insights from SAP-generated tabular reports, enhancing data-driven decision-making. • Streamlined report generation through the use of Llama Hub Tools for data visualization, reducing the time required by 60% and increasing efficiency in report production. LSTM Model-based Time Series Forecasting of Sales and Consumption • Devised a demand forecasting system in the supply chain, significantly enhancing the accuracy in predicting market demand and consumption patterns. • Conducted meticulous data preprocessing, including normalization, sequence generation, and handling of temporal data to ensure high-quality inputs for the forecasting model • Adopted advanced deep learning techniques to unravel complex demand patterns, resulting in a 40% improvement in forecast accuracy and more reliable demand predictions.
	Software Development Engineer \| AJIO Inc. Mumbai, India Nov '20 - Jun '22 Comprehensive B2B Tax Calculation and Data Processing System • Crafted Java-Spring Boot micro-services with Apache Kafka Streaming, ensuring seamless invoice data extraction • Automatized feature extraction from invoices by using Azure Cognitive Services, from over 10,000 invoices per day and deployed a PySpark-based data pipeline to store the extracted features in Azure Data Lake Storage (ADLS) • Orchestrated fault-tolerant micro-services deployment on Kubernetes, elevating scalability, and resource efficiency • Integrated the micro-service system into Ajio’s Service Mesh for catering to 9 internal teams ensuring widespread adoption
	Web Developer Intern \| Things Go Social Delhi, India Aug '20 - Nov '20 • Updated an existing system from jQuery and plain HTML to a more robust MERN stack-based web application, contributing to API development for various endpoints. • Built a single-page web application using the MERN stack with Redux, Webpack, and Babel to enhance functionality and user experience. • Created micro-frontends to reduce server-side load, optimizing the performance and scalability of the front-end components.

	Document Data Analyzer PyTorch, Hugging Face, Python, OpenAI, GenerativeAI, Retreival \| Nov. 2023 - Dec. 2023 [code][demo] • Spearheaded building RAG (Retrieval Augmented Generation) chatbot powered by vector databases and LLMs • Leveraged HuggingFace Sentence Transformer “all-MiniLM-L6-V2” for embeddings and FAISS Similarity Search • Assimilated OpenAI’s API, LangChain framework and FAISS vector store for orchestrating the pipeline for querying model
	LLM-Powered SQL DB agent LangChain, OpenAI, LLM, Python, SQL, APIs, SQLAlchemy \| Feb. 2024 - Apr. 2024 [code][demo] • Developed a foundational NL2SQL model using LangChain that translates natural language queries into SQL commands, allowing users to interact with databases without needing SQL knowledge. • Implemented rephrasing techniques using LangChain’s prompt templates to convert SQL results into clear, natural language responses, improving user understanding of the data. • Utilized dynamic few-shot example and relevant table selection to tailor model responses based on query context, ensuring accurate and efficient handling of various user queries. • Integrated memory capabilities to retain conversational context, enabling the model to accurately address follow-up queries and provide a seamless user experience.
	LSTM Stock Predictor LSTM, pandas, time series forecasting \| Jun. 2023 - Jul. 2023 [code] • Established a time series forecasting model using a windowed approach and LSTM neural network to predict stock prices, transforming pandas DataFrame data into a format suitable for LSTM training. • Defined the LSTM model architecture with an input layer, an LSTM layer with 64 units, and two dense layers with ReLU activation, compiled using mean squared error loss and the Adam optimizer. • Conducted model training over 100 epochs, fitting the model on training data and validating on a separate validation set, ensuring robust performance. • Visualized model predictions against actual observations for training and testing datasets, effectively capturing temporal dependencies in historical closing prices to enhance forecasting accuracy.
	Twitter Data ETL automation system AWS S3, AWS EC2, AWS-managed Apache Airflow, DAG, CSV, Data Validation frameworks \| Sep. 2023 - Oct. 2023 • Established ETL frameworks on the AWS platform utilizing services such as S3 and EC2 to automate the creation of Twitter data CSV files. • Initiated data pipelines using AWS-managed Apache Airflow by creating DAGs, automating end-to-end ETL jobs for efficient data processing. • Acquired proficiency in data validation and quality check frameworks to handle real-time data, ensuring 100% data quality for consumption in the form of CSV files deployed on an S3 bucket.
	Ticket Generating System Next.js, Tailwind CSS, MongoDB, Express.js \| May. 2023 - Aug. 2023 [Code] • Constructed a Next.js system for allocating system tickets to resolve issues based on severity and priority, ensuring efficient issue management. • Devised an MVC architecture with a ticket schema stored in a NoSQL format in MongoDB using Atlas, providing a scalable and flexible data storage solution. • Implemented all CRUD operations with APIs written in an Express.js system, leveraging Next.js routing capabilities to manage the ticket lifecycle effectively.