Paddy Disease Detection and Classification Using Computer Vision Techniques
Bimarsha Khanal et al. (2024). "Paddy Disease Detection and Classification Using Computer Vision Techniques." arXiv:2412.05996.
CV
Education
- B.E. in Computer Engineering, Tribhuvan University, Institute of Engineering, Western Region Campus, 2019–2024
- Coursework: Software Engineering, Operating Systems, DSA, AI, DBMS, Big Data Technologies, Distributed Systems
- Awarded a full undergraduate scholarship for ranking in the top 3% of the IOE entrance examination
- Fusemachines AI Fellowship 2023: Year-long Microdegree program in Artificial Intelligence
Research Interests
- Multimodal LLMs and AI Agents
- Trustworthy AI, Privacy Preserving, and Secure AI Systems
Work experience
- March 2024 – Present: Machine Learning Engineer
- Fusemachines, Kathmandu, Nepal
- Built and delivered AI-driven systems, including document intelligence pipelines for financial workflows and research on multi-modal LLM-powered interview agent for assisting hiring managers throughout recruitment process.
- Collaborated with cross-functional teams to integrate ML solutions into production and communicate technical concepts to both technical and non-technical stakeholders.
- March 2026 – Present: Teaching Assistant
- Fusemachines AI Fellowship, Nepal
- Conducted sessions in a six-month AI fellowship program for students and working professionals on deep learning fundamentals and real-world applications.
- Mentored fellows through project ideation, implementation, and evaluation; collaborated on session planning and curriculum materials.
Skills
- Machine Learning: Deep Learning, Transformers, Preference Optimization, RAG, AI Agents, Multimodal Learning, SFT, DPO
- Frameworks: PyTorch, Hugging Face Transformers, Scikit-learn, OpenCV, ONNX
- Programming: Python, C++, JavaScript, Bash, SQL
- Systems & Infrastructure: Linux, Docker, AWS, HPC Environments, Vector and Graph Databases
Projects
- Document Extraction Engine: end-to-end extraction from KYC, financial, and rental documents (classification, deep homography alignment, custom handwriting recognition, spelling correction); evaluated DiT, Table Transformers, PaliGemma, Qwen-VL, LLM-based methods; RAG + agents with active learning for validation, discrepancy detection, and deal-specific Q&A; scalable parallel pipelines handling thousands of docs daily.
- Audio-LLM Research: multimodal interview system with Qwen2-Audio + LoRA/DPO; synthetic speech/conversation and preference data generation with voice cloning; real-time VAD and speaker diarization pipelines.
Publications
Talks
Engineering AI Systems
Fellowship session at Fusemachines AI Fellowship, Kathmandu, Nepal
Applied AI
Fellowship session at Fusemachines AI Fellowship, Kathmandu, Nepal
Certifications
- Claude Certified Architect - Foundations, Anthropic, August 2026
Awards
- Delta Hackathon (2024): Public Littering Detection and Waste Management System, First Runner-Up
- Leapfrog Revamp Hackathon (2023): Personalized News Portal, First Runner-Up
- Ethos Hack: Cultural Hackathon: Cultural Entities Detection and Information Center, Winner
- Vertex Hacks Chapter Alpha (2023): Remote Tourism Booking System, Finalist
- Techparva Datathon (2022): Election Outcomes of Nepal, First Runner-Up
Teaching & Mentoring
Service and leadership
- Leapfrog Student Partner (2023), Leapfrog Technology: 9-month program; mentored undergraduates on career development and technical skills.
- Python and Machine Learning Mentor (2023, 2024), Innovative Computer Engineering Society: month-long Call for Enthusiast program guiding junior students in ML.