Hi, I'm Moulik
Introduction
Overview.
Data Scientist
Software Engineer
Analyst
AI/ML Engineer
Product Manager
What I have done so far
Extracurricular Activities.
Head of Coding Department
DJS PhoenixMarch 2020 - November 2021
- • Created an autonomous drone system with the help of Python and a OpenCV for my team’s Immediate Aid Assistance Drone (IAAD) which helped in healthcare services and became the finalist in the Indian Institute of Project Technology Competition.
- • Managed a team of 10 developers in designing and implementing an autonomous object tracking system, overseeing coding tasks and ensuring project milestones were met efficiently.
Marketing Associate
DJS BeatsJanuary 2020 - August 2020
- • Initiated contact with different companies to seek sponsorship for the events organised by our Student Forum.
- • Coordinated data tracking, events, interviews, and managed finances, including budget monitoring.
My work
Projects.
These projects showcase my practical skills and experience, each with descriptions and links to code repositories and live demos. They demonstrate my ability to handle complex challenges, adapt to different technologies, and oversee projects from start to finish.
Cross-Domain Adaptation through Soft Prompt Tuning in Low-Parameter Language Models
•Optimized cross-domain adaptation for low-parameter language models using soft prompt tuning (e.g. prefix tuning) on T5-large, achieving over 90% accuracy in sentiment analysis across multiple domains with less than 1% of model parameters trained. •Improved few-shot learning in resource-constrained models through techniques like LoRA and soft prompting, enhancing sentiment analysis accuracy while reducing the need for extensive model fine-tuning
#python
#pandas
#LoRA
#LLMs
#Prompt Tuning
LLM-Powered Financial News Sentiment Trader
•Created LLMSentiments Trader with DistilBERT and FinBERT for real-time NER and sentiment analysis on financial news, supporting investment decisions for hedge funds. Integrated Point72’s CSP library, improving stream processing efficiency. •Implemented a sentiment-based trading strategy that yielded a 0.21% return with a sharpe ratio of 0.84, outperforming traditional methods by analyzing a 20GB dataset of global news to track sentiment-driven stock movements.
#python
#pandas
#TensorFlow
#BERT
#UNIX
Smart Interactive Marketing
• Built a CNN model for fashion product classification using the DeepFashion Dataset (800k images), achieving 88% accuracy and integrating it into a React.js application for real-time recommendations, which boosted product engagement by 30%. • Created an Object Detection System leveraging ResNet50 and VGG-16, achieving 62% accuracy and enhancing product recommendations by 20% through effective image analysis and feature extraction.
#Python
#TensorFlow
#React.js
#CNN
#Node.js
Smart Mart
• Developed a Smart Shopping Basket using YOLOv3 for accurate product identification 81% accuracy, connecting it with an e-commerce web platform via React.js and Flask to streamline user interaction. • Deployed a real-time recommendation system that clustered user behavior and preferences, enhancing shopping experiences by 30% through personalized product suggestions tailored to user segments.
#Python
#TensorFlow
#YOLOv4
#Flask
#React.js
Speech Emotion Recognition System
• Engineered a CNN-based speech-emotion recognition system with 82% accuracy using TensorFlow, MFCC features, and Wav2Vec2 from HuggingFace, classifying emotions like happiness, sadness, anger, and fear from speech. • Implemented a real-time emotion classification via Gradio app and Flask API, increasing user engagement by 18%, leveraging advanced signal processing techniques like MFCCs and Fourier Transform.
#Python
#TensorFlow
#Gradio
#Flask
#Hugging Face
What I have learned so far
Education.
My academic journey has equipped me with both theoretical knowledge and practical skills, preparing me for real-world challenges in technology and innovation.
- •Relevant Coursework: Machine Learning, Natural Language Processing, Computer System Architecture
2019 - 2023
CGPA: 8/10
Bachelor of Technology in Computer Engineering
Dwarkadas J. Sanghvi College of Engineering- •Relevant Coursework: Data Structures, Algorithms, Machine Learning, Deep Learning, Natural Language Processing, Computer Vision, Database Management Systems
2017 - 2019
Percentile: 93.2%
Higher Secondary Certificate (HSC)
Prakash College of Commerce and Science- •Specialized in Science with focus on Physics, Chemistry, and Mathematics
- •Represented at National Level in Throwball and participated in Inter School and State Level Competitions
- •Represented at School Level in Volleyball participated in Inter School Competitions
Achievements.
Recognition & Awards
Led the development of an innovative autonomous drone system for healthcare services, integrating AI-driven navigation and medical supply delivery capabilities. The project advanced to the finals of the prestigious IIPT competition, demonstrating technical excellence in combining healthcare logistics with autonomous systems.
Successfully completed The Math Company's competitive aptitude assessment, leading to a direct job offer. Demonstrated exceptional problem-solving and analytical skills throughout the competition.
My skills
Technologies.
Frameworks
Get in touch