AI/ML Engineer and Computer Engineering graduate (Kathmandu University). I work across NLP, Computer Vision, and MLOps, and I like taking ideas the whole way — from a research paper to a system that actually runs in production.

First-author on multiple publications, contributor to a national platform serving 300K+ users, and co-founder of two AI products.

🌐 sushanadhikari.com.np  ·  🎓 Google Scholar  ·  💼 LinkedIn  ·  ✉️ sushan.adhikari2060@gmail.com

• Research → production. Low-resource NLP, LLM fine-tuning + RAG, and CV inference — published first, then deployed.
• Computer vision at scale. Optimized 5+ models on NVIDIA Triton and benchmarked DeepStream + GStreamer for real-time multi-camera pipelines (Computer Vision Intern → Junior MLOps Developer at KyraWorks).
• MLOps. CI/CD for model deployment with Docker, AWS, EC2, and Terraform.
• Work that ships. Integrated anti-spoofing liveness checks into Nepal's national e-pension platform (300K+ users), and built an ML job-recommendation pipeline across 50+ categories with 150+ global volunteers at Mercuri.world.

• Directional Asymmetry in Low-Resource Legal Machine Translation: A Nepali–English Case Study — ICAIL 2026
• Enhancing Ethical Reasoning in Tiny LLMs via Fine-Tuning and Multi-Agent Consensus — KUSET, Vol. 18 No. 2 (2024)
• Detecting Image Forgeries and Deepfakes: A Comparative Study of CNN and Transformer Models — KUSET, Vol. 18 No. 2 (2024)
• Physics-Informed Data Augmentation for Sediment Concentration Prediction in Himalayan Rivers — IJETID, Vol. 2 No. 2 (2026)
• RAG for Theoretical CS Education (under review) · explainability for hybrid CNN-Transformer document layout analysis & automated math-animation generation (ongoing)

→ Full list on Google Scholar.

• Nurvexa — RAG-powered nursing-licensing exam prep (co-founder)
• PrepGraduate — adaptive GRE practice + score analytics (co-founder)

• Cosmira — interactive solar-system explorer · NASA Space Apps People's Choice winner
• Agni-Netra — IoT + computer-vision fire & smoke detection on Raspberry Pi
• MastiskaTrack — NLP + LLM mental-health assessment · AI Crusade winner
• AlgoRAG — retrieval-augmented tutor for algorithms & complexity theory
• XAI_DLA — explainability for hybrid CNN-Transformer document layout analysis
• liveness-detection — full-stack anti-spoofing liveness checks
• CrowdFunding-Blockchain — Web3 crowdfunding with on-chain accountability
• Manim_Visualize — LLM-generated mathematical animations

Python PyTorch scikit-learn Hugging Face LangChain FastAPI Django React
NVIDIA Triton DeepStream GStreamer YOLOv8 OpenCV
AWS Docker Terraform Airflow PySpark PostgreSQL

Machine Learning Specialization — Andrew Ng (Coursera) · Apache Airflow: DAG Authoring & Fundamentals (Astronomer) · DataCamp (ML, PySpark, SQL) · Embeddings for NLP (OpenHPI)

Hult Prize OnCampus Winner — led the computer-vision pipeline for Dr.Fish (fish-disease classification) and represented Nepal at the Regional Summit in Bangkok · NASA Space Apps People's Choice · ICT Awards 2024 Finalist · AI Crusade Winner (Health Track) · Mahatma Gandhi Scholarship

Internal Relations & Operations Lead for KU HackFest 2024 (~200 team applications); active in KUCC, Leo Club, and Rotaract. Away from the keyboard: piano, reading, and volunteer engineering for social-impact teams. Languages: English, Nepali, Hindi, and beginner German.

_{B.Eng. Computer Engineering, Kathmandu University · exchange semester at IIT Palakkad · based in Nepal}