AI/ML & Data Engineer | MS in Computer Science @ Arizona State University (4.0 GPA)
I build intelligent systems at the intersection of large language models, retrieval, and large-scale data engineering. Previously spent ~5 years at Dell Technologies engineering real-time fraud-detection pipelines and semantic retrieval systems that led to a granted patent in language-agnostic hybrid search.
Building LLMs, RAG systems, and applied ML - from local multimodal PDF chat to real-time object detection on edge hardware.
- 🔭 Currently building with LLMs, RAG, and multimodal pipelines
- 🎓 MS CS @ ASU (Aug 2025 – May 2027)
- 📄 Patent holder — Distributed Hybrid Search for Language-Agnostic Retrieval
- 📫 Reach me at daripa.ushnesha9701@gmail.com
AI / ML
Data & Systems
Cloud & Tools
🛰️ AI-Powered Autonomous Trash Interceptor Real-time object detection and trajectory prediction on a Raspberry Pi with an ArduCAM ToF depth sensor (45%+ catch accuracy). Built a dual prediction pipeline pairing a physics-based ballistic model with a RandomForest variant trained on 500+ trajectories.
🐦 BirdCLEF+ 2026 (Kaggle) Pretrained and fine-tuned EfficientNet backbones to identify 234 wildlife species from audio recordings of Brazil's Pantanal wetlands. Converted audio to mel-spectrograms, trained with augmentation and AUC-optimized loss, and ensembled segment-level predictions with blended ProtoSSM, SED and BirdNet achieving best score of 0.944.
📄 Multimodal RAG Chat with PDFs (Local LLM) End-to-end multimodal RAG system supporting text, tables, and images from PDFs — fully local and privacy-preserving using Ollama, vector embeddings, and Streamlit.
