Transforming Data into Insights | IIT Kharagpur Graduate
Passionate about turning data into actionable insights
I'm a dedicated Data Scientist and AI Enthusiast with a strong foundation in machine learning, deep learning, and natural language processing. As a graduate from IIT Kharagpur, I combine technical expertise with creative problem-solving to deliver impactful solutions.
My journey in data science has led me to work on diverse projects ranging from health data processing platforms to industrial motor fault detection systems. I'm passionate about leveraging AI to solve real-world problems and drive meaningful change.
Indian Institute of Technology, Kharagpur
CGPA: 7.74/10 | 2024
D.A.V Public School, Patna
86.40% | 2019
D.A.V Public School, Patna
9.6/10 | 2017
Showcasing my technical expertise and innovation
My professional journey in data science
Apr 2024 - Ongoing
* Product Development: Engineered a production-ready, AI-powered health
platform using FlutterFlow for the frontend and Supabase for the backend database
architecture.
* AI Coaching System: Implemented a sophisticated AI health assistant by integrating
OpenAI Assistant APIs and building a custom RAG (Retrieval-Augmented Generation)
pipeline.
* Data-Driven Insights: Developed logic to process daily activity logs and complex
medical
health reports to generate personalized, actionable health recommendations for
users.
* Architecture: Designed the system to serve as a 24/7 personal health assistant,
ensuring
low-latency responses and secure handling of sensitive health data within the Supabase
ecosystem.
Apr 2023 - Jul 2023
* Ensemble Modeling: Developed a high-performance FAQ classification model by leveraging
advanced ensemble techniques to categorize edtech queries.
* NLP Pipeline & Engineering: Engineered a dataset of 6,000+ edtech FAQs, implementing
comprehensive NLP preprocessing and utilizing Transformer architectures for deep
semantic understanding.
* Feature Extraction & Analysis: Employed TF-IDF vectorization and initially trained a
Random Forest Classifier, achieving 74% accuracy in sentiment analysis of user
inquiries.
* Optimization & LDA: Achieved a peak accuracy of 94% through rigorous hyperparameter
fine-tuning and the integration of Latent Dirichlet Allocation (LDA) for sophisticated
topic modeling.
My technical toolkit
Recognition and certifications
Earned prestigious certification with focus on image classification
2024IBM-endorsed certification in 'Excel Basics for Data Analysis'
20241st rank in Akhil Bhartiya Rashtra Bhasha Examination
2023Completed 'The Data Scientist's Toolbox' course on Coursera
2023Top honors in nationwide Canvas Competition
2023Let's discuss opportunities and collaborations
deeps98kumar@gmail.com
India