Ph.D. Student in Computer Science · Vanderbilt University
I work at the intersection of machine learning, data engineering, and cloud-native software systems.
- Portfolio: bibekdhungana.com
- Resume: bibekdhungana.com/resume
- Certifications: credly.com/users/dhunganabibek/badges
- LinkedIn: linkedin.com/in/dhunganabibek
Four years of industry experience shipping systems at scale:
- ETL workflows reconciling 11M+ records on AWS Glue
- Kafka streaming pipelines sustained at 99.9% uptime
- Full-stack platforms serving millions of requests weekly
- LLM & retrieval systems for semantic search and research automation
-
ResearchTeam: LLM platform matching researchers to 80,000+ federal grants via dense semantic retrieval (MRR 0.74 vs. 0.54).
-
StreamCab: Real-time fare-prediction pipeline on Kafka and Spark; XGBoost reaches 3.42% MAPE (78% lower error).
-
Quantum Adaptive Self-Attention (QASA): Hybrid quantum-classical Transformer (PyTorch, PennyLane); matches a classical baseline (R² 0.88 vs. 0.90) with ~6% fewer parameters.
-
RAG: The Philosophical Computer: Fully local retrieval-augmented generation (LangChain, ChromaDB, Ollama) with zero external API calls.
-
Apparel Classifier: Fashion-MNIST model served via FastAPI and React with live camera inference.
| Area | Tools |
|---|---|
| ML / AI | PyTorch Hugging Face scikit-learn LLMs RAG LangGraph MLflow |
| Data Engineering | Spark Kafka Airflow AWS Glue Pandas NumPy |
| Cloud / MLOps | SageMaker Lambda Docker Kubernetes Terraform GitHub Actions |
| Databases | PostgreSQL Snowflake Redshift BigQuery MongoDB DynamoDB Pinecone |
| Languages | Python SQL TypeScript Rust Java C# |
| Web | React Next.js Node.js FastAPI .NET Core Spring Boot |