- 🔭 Currently working on transformer architectures and building them from scratch
- 🌱 Deep diving into MLOps best practices and automated ML pipelines
- 🎓 Graduate from NIT Warangal with a background in Remote Sensing and GIS with focus on GeoAI
- 🌍 Based in Delhi, India
- 💬 Ask me about LLMs, CNNs or Geospatial deep learning
- ⚡ Fun fact: I love implementing papers from scratch to truly understand the architecture
Master's thesis project comparing deep learning algorithms for Land Use Land Cover classification using 1D CNNs. Applied to real-world geospatial datasets.
Tech: Deep Learning Geospatial Analysis Remote Sensing 1D CNN
Experimenting with end-to-end MLOps pipelines, CI/CD for ML models, model versioning and deployment best practices.
Tech: MLOps Docker CI/CD Model Deployment
Built transformer and Vision Transformer (ViT) architectures from scratch in PyTorch. Deep dive into attention mechanisms, positional encodings and patch embeddings.
Tech: PyTorch Transformers Vision Transformer (ViT) NLP
Fine-tuning Llama 3 on the Dolly dataset using Low-Rank Adaptation (LoRA) for parameter-efficient training. Exploring efficient fine-tuning techniques for large language models.
Tech: Llama 3 Weights and Biases LoRA
Some interesting projects and resources I've been exploring:
- ucam-eo/tessera - TESSERA is a GeoAI foundation model that can process time-series satellite imagery
- VMarsocci/pangaea-bench - A benchmark for evaluation of Geospatial Foundation Models
- DataTalksClub/mlops-zoomcamp - MLOps course from DataTalks.Club
- 🔥 Exploring GeoAI foundation models
- 📚 Building intuition on efficient training pipelines for transformer models
- 🧪 Experimenting with Retrieval-Augmented Generation (RAG) systems
I'm always open to collaborating on interesting AI/ML projects, discussing research ideas or just chatting about the latest in deep learning!