🚀 Megh Dave

📍 Mumbai, Maharashtra
📧 [email protected]
📞 9222056520
🔗 GitHub: Megh06

---

🏆 About Me

I am a Computer Science (Data Science) undergraduate with a strong focus on AI/ML, Deep Learning, NLP, and Multimodal Retrieval Systems. I enjoy building end-to-end AI solutions—from research and modeling to deployment—while keeping solutions scalable, interpretable, and business-friendly.

My interests lie at the intersection of Machine Learning research, Generative AI, and real-world applications, with hands-on experience in PINNs, time-series modeling, multimodal RAG systems, and AI-powered learning platforms.

---

🎓 Education

Dwarkadas J. Sanghvi College of Engineering
B.Tech in Computer Science (Data Science)
Honors in Computational Finance
📊 CGPA: 8.93
📅 Aug 2023 – Present
📍 Mumbai, Maharashtra

---

🧠 Experience

🔬 ML Research Intern — IIT Bombay

📅 Jun 2025 – Present

• Identified modeling gaps in existing plant transpiration studies and proposed an inverse Physics-Informed Neural Network (PINN) framework to estimate latent plant parameters.
• Integrated LSTM-inspired temporal encodings for better representation of time-dependent dynamics.
• Applied TimeGAN-based noise injection to improve robustness and generalization under sparse and noisy observations.
• Worked on research-grade modeling, focusing on interpretability and physical consistency.

---

🛠️ Skills

💻 Programming Languages

• Python, HTML, CSS

⚙️ Frameworks & Tools

• FastAPI, Flask, Streamlit
• React, Bootstrap, Tailwind CSS
• Docker

🧠 AI & Machine Learning

• Supervised Learning & Feature Engineering
• Deep Learning (CNNs, RNNs, LSTMs)
• Physics-Informed Neural Networks (PINNs)
• Generative AI & TimeGANs
• NLP, Information Retrieval, RAG Pipelines

🗄️ Databases & Cloud

• MySQL, ChromaDB
• Cloudinary
• Git, GitHub

---

🌟 Projects

📚 OOPS!IDidntStudy — Last-Minute AI Study Platform

PanicNotes

• AI-generated consolidated notes from PDF, DOCX, PPTX, and Word files.
• Merges multi-source content into a single streamlined study document.

CramBot & Quizzard

• CramBot answers questions using user-uploaded notes.
• Quizzard generates adaptive quizzes for exam preparation.

Tech Stack:

Python · Deep Learning · NLP · React · FastAPI · Vercel · Streamlit

---

🧠 JigyasaAI — Multimodal Retrieval-Based Chatbot

• Users upload PDFs and ask contextual questions.
• Retrieves relevant images + text answers, unlike chatbots focused only on image generation.
• Uses CLIP-based embeddings for multimodal alignment.

Tech Stack:

Python · CLIP Models · ChromaDB · Cloudinary · APIs · Streamlit

---

🌿 Potato Leaf Disease Detection

• Built a deep learning classifier (99%+ accuracy) for real-time potato leaf disease diagnosis.
• Supports user-uploaded images with fast inference.

Tech Stack:

Python · TensorFlow · FastAPI · Streamlit

---

🏁 Hackathons & Achievements

• 🥈 2nd Runner-Up, Co-Code Hackathon (Hybrid Mode), DJ Sanghvi College
• 🥈 2nd Runner-Up, DJS Sanshodhan 2025

---

📜 Certifications

• AWS Academy Graduate – Machine Learning Foundations
• AWS Academy Graduate – ML with NLP Foundations

---

🌐 Co-Curricular Activities

DJS S4DS — Marketing Associate

📅 Aug 2024 – Present

• Identified and engaged sponsors to secure funding and partnerships.

DJS Nova — Publicity Associate

📅 Sep 2024 – Present

• Managed social media and offline publicity to maximize event engagement.

---

🚀 Mission

To bridge research and real-world impact by building robust, scalable, and intelligent AI systems—from theory-driven models to production-ready applications.

---

⭐ If you find my work interesting, feel free to explore my repositories or connect with me!