ETH Zurich MSc Data Science | Software Engineering | Computer Vision | 3D Reconstruction | ML Systems
I build production-minded AI systems with research depth across computer vision, 3D reconstruction, retrieval, edge AI, and data-heavy engineering.
I am an MSc Data Science student at ETH Zurich, currently working across computer vision, 3D reconstruction, autonomous perception, and applied AI engineering. My experience includes research pipelines, edge deployment, MLOps, DevOps automation, data infrastructure, and open-source contributions.
- Researching robust indoor 3D reconstruction at ETH Zurich's Computer Vision and Geometry Group.
- Built monocular depth-estimation systems for autonomous truck perception at Caterpillar.
- Built billion-record ETL and observability pipelines at Akamai.
- Led federated learning and GAN-based research projects with published and award-winning outcomes.
Computer Vision 3D Reconstruction Edge AI
Retrieval + RAG Multimodal VLMs MLOps
Distributed Systems Data Engineering Production AI
Deployment Infrastructure
| Project | What it does | Stack |
|---|---|---|
| Relational Reasoning in VLMs | Studies cross-modal attention, shortcut learning, activation patching, probing, and relational reasoning in Qwen-based VLMs. | PyTorch, Qwen, CLEVR, ChartQA |
| Federated IDS with DBNs | Privacy-preserving WiFi intrusion detection with federated training on 300k+ sensor records and 98% accuracy. | PyTorch, gRPC, Federated Learning |
| Doc-Chat | PDF chatbot with ingestion, semantic chunking, vector indexing, grounded Q&A, summarization, and quiz generation. | Flask, RAG, ChromaDB |
| Blood Connect | Blood donation management system with donor records, requests, and PostgreSQL-backed workflows. | PostgreSQL, Web App |
| Lyric Lullaby | Music streaming app with authentication, playlist management, search, and relational data modeling. | Flask, SQLAlchemy |
| Handwriting Letter Classification | Feed-forward neural network for handwritten character recognition with 90%+ EMNIST accuracy. | Neural Networks, EMNIST |
| Role | Organization | Highlights |
|---|---|---|
| Research Student | ETH Zurich CVG | 3D reconstruction for in-the-wild indoor video, HouseTour, COLMAP, DUSt3R, MASt3R, DAV3. |
| Software Engineer Intern | Caterpillar Inc. | Monocular depth estimation for autonomous truck perception, 50k+ images, edge AI optimization. |
| Software Engineer Intern | Akamai Technologies | Jenkins ETL pipelines, Hive, ClickHouse, Grafana, billion-record daily processing. |
| Research Intern | Samsung R&D Institute PRISM | Conditional GANs for video sequence generation, 0.97 SSIM, team leadership. |
- Accepted Springer 2025 paper:
FEDDBN-IDSfor federated intrusion detection on heterogeneous WiFi networks. - LeetCode Knight, top 2.5% globally with 2020+ rating.
- Winner, Internal Smart India Hackathon 2023.
- 2nd place, Innovision24 Hackathon among 48 teams.
- A. John Wilson Endowment Prize for highest Mathematics score across departments.
- Samsung PRISM Excellence Certificate.
I am always happy to talk about computer vision, ML systems, 3D reconstruction, research ideas, and engineering roles.
