Introduction

Results-driven Data Scientist and Deep Learning researcher with an exceptional track record of optimizing operations and reducing costs.
Proficient in Python, PyTorch, SQL, PYOMO, Gurobi, CPLEX, C/C++, MATLAB programming.
Expertise in real-time scheduling, resulting in multiple high-quality publications and dynamic presentations.

Education

Ph.D., Electrical & Computer Engineering	Forida International University	May 2022
M.E., Electronics (Digital Systems)	Savitribai Phule Pune University	Sep 2013
B.E., Electronics & Telecommunication	Savitribai Phule Pune University	Aug 2010

Work Experience

Data Scientist | Crowley, Jacksonville, FL, US | June 2022 – Current

Demonstrated exceptional proficiency in prescriptive analytics, particularly in optimization problem formulation, resulting in a noteworthy 3%–4% cost savings for the strategic repositioning of the empty containers.
Demonstrated rapid learning capabilities by swiftly acquiring expertise in natural language processing, leading to the development of a private and secure LLM-based chatbot prototype through adept rapid prototyping.
Leveraged SQL skills to fetch, join, clean and preprocess complex data from multiple sources, implementing non-periodic time series predictive model training/deployment contributing to 8% increase in bookings count, customer experience and retention.
Brainstorming to share expert advice, statistical analysis to recognize patterns, data interpretation, and advanced machine learning resulted in the building and training of the prototype for predicting tugboat engine maintenance.
Ability to work in a fast-paced environment and skills to handle poor quality data resulted in rapidly debugging and scaling existing overbooking prediction model.
Demonstrated strong communication skills and applied computer vision expertise in the implementation of port gate automation, effectively capturing chassis numbers attached to trucks.

Graduate Researcher | Florida International University, Miami, FL, US | August 2018 – April 2022

Skills in developing deterministic and testable code with expertise in real-time system design as evidenced by novel temperature estimation framework for vehicle Electronic Control Units (ECUs).
Ability to review others work and solid advanced mathematics fundamentals invented Deep Neural Network sensitivity analysis method for accuracy maximization by 0.5 – 4.2% for standard architectures and datasets.
Innovative thinking, knowledge of statistics, and optimization skills demonstrated by thermal-aware mapping of Deep Neural Network neurons on 3D memory banks for accuracy maximization in the range 0.18 – 47.91%.

Technical Consultant | Sujlam Electronics, Baramati, MH, India | March 2016 – December 2017

Sensor interface debugging skills and experience of embedded firmware development led to the detection of dry-run problems, guiding the high performance and flexible control circuitry development.
Excellent communication and collaborative team player skills, and experience of writing peripheral drivers as demonstrated by 7% lower cost electrical pump protection product than competitors.

Assistant Professor | Savitribai Phule Pune University, Baramati, MH, India | July 2013 – July 2018

Experience of mapping Image Processing algorithms on FPGAs resulted in a research project grant of INR 170K.
Successfully elevated student engagement and academic performance in VLSI Design over a five-year period, driving a remarkable 21% increase in pass rates and an 18% improvement in maximum marks.
Creative thinking and experience of development using Real-Time Operating System kernel with micro-controllers resulted in new lab instructions for Advanced Processors course.

Projects

Stock Price Forecasting using Temporal Fusion Transformer - In this project, I employ advanced technique to predict stock prices accurately over time. Through temporal fusion modeling, I leverage temporal patterns and external factors to generate reliable forecasts, aiding in informed decision-making in the financial markets. The DOW JONES stocks prices for the next 5 days are predicted with a MAPE of 0.16. GitHub link

Fundamental Analysis of the Stocks listed in NSE, India - Explore my educational and enjoyable project, delving into the financial markets to understand stocks listed on the National Stock Exchange (NSE). Through comprehensive fundamental analysis techniques, I aim to uncover insights into company performance, utilizing financial statements and key indicators to evaluate growth prospects and profitability. Access the web app to visually explore stock fundamentals for free (except Finance sector companies including Banks and NBFCs) - Web app link. GitHub link

Generative AI-Based Data Analysis for Evaluation of Variations in XPS Characteristics of MoS - I developed a Large Language Model (LLM)-based assistant to aid my friend’s research on MoS2 material. The XPS analysis for conformance of MoS2 revealed that there is significant variation in the Mo3d and S2p plots. The LLM assistant streamlines processes and enhances accuracy in XPS analysis. More details are available in the research paper here.

Homeopathy Assistant: Leveraging Large Language Models for Remedies - This project utilizes Large Language Model (LLM) to streamline the process of prescribing remedies, revolutionizing how homeopathy doctors address patient health concerns. With a focus on the holistic approach of homeopathy, it aims to provide efficient solutions leveraging advanced language processing tailored to individual symptoms and conditions, saving time for practitioners. GitHub link

Fine-tuning Microsoft Phi-2 Language Model for Sentiment Analysis - I demonstrate the fine-tuning of the Microsoft open-source Phi-2 Language Model for sentiment analysis on financial and economic data. Using the Simple Fine-tuning Trainer (SFTTrainer) with the Parameter-Efficient Fine-Tuning (PEFT) method, I efficiently refine the model’s parameters, optimizing computational resources while maintaining accuracy. GitHub link

Emotion Recognition from Voice using NLP - This innovative project employs Natural Language Processing (NLP) to analyze emotions from voice recordings by leveraging Google’s speech-to-text API, enabling accurate emotion decoding. Through advanced machine learning algorithms based on Ekman’s six basic emotions model, including happiness, sadness, anger, fear, disgust, and surprise, the system interprets emotions intuitively. With applications in mental health support, customer feedback analysis, and interactive voice systems, recognizing emotions from voice enhances interaction quality, enables personalized responses, and offers insights for decision-making.

Multi-Modal Shortest Path Optimization for Goods distribution - Here I developed a shortest path algorithm considering transportation costs via road, rail, air, and ocean, coupled with storage and shortage costs, to determine the most cost-effective route for goods distribution in a complex retail network. The linear mathematical model is developed using PYOMO framework and Gurobi solver is used for achieving optimal solution.

Capacitated Vehicle Routing Problem - In this project, I tackle the classical operations research problem. The goal is to design optimal delivery routes and assign vehicles to them in order to minimize the total distance traveled. To accomplish this task, I employ the PYOMO framework, a powerful optimization modeling tool, along with the open-source GLPK solver. GitHub link

Technical Skill Set

Python, PyTorch, Dataiku, SQL, PowerBI, Snowflake, Hugging Face, LangChain, VectorDBs, Pyomo, Matplotlib, Scikit-learn, Pandas, CPLEX, Gurobi, Linux, Github, AWS, Azure OpenAI and Form Recognizer, C/C++, Embedded C, Matlab.

Awards and Recognition

Crowley TOPS awards for displaying high performance value in September 2022, August 2023 and Jan 2024.
Featured in Crowley’s blog post “Data-driven Decision-making in the Supply Chain.”
Dissertation Year Fellowship award for Fall 2021 and Spring 2022 semesters by Florida International University.

Extra Curricular Activities

Steering committee member of Asian American Pacific Islander Business resource group at Crowley.
Volunteered for the “Real Men Read” event at San Jose Elementary School, Jacksonville, FL, United States, Nov 2023.
Volunteered for “International Yoga Day” event at Jacksonville Beach, FL, United States, Jun 2023.
Invited speaker for presentation on “Higher Education and Career Opportunities in the United States” in Tuljaram Chaturchand College, Baramati, MH, India, Feb 2023.
Judged abstracts for the “Graduate Students Appreciation Week,” assessing research methods, results and limitations at Florida International University FL, USA, Feb 2021,
Invited speaker for presentation on “Emerging Trends and Technology in Electronics” in National webinar at Tuljaram Chaturchand College, Baramati, MH, India, Feb 2021.
Presented research paper in “2020 IEEE 38th International Conference on Computer Design (ICCD),” Hartford, CT, USA, Oct 2020.
Judged poster presentations for the event “CURFIU 2019,” assessing visual appeal, content clarity, and presentation effectiveness at Florida International University, FL, USA, Apr 2019.

Hobbies and Interests

Watch live Cricket matches
When I am not in front of the computer, either I am playing with my kids or practicing yoga and meditation.
Volunteer for the Art of Living organization to spread the health and happiness in the community.

Last updated: 7-April-2024