A bit about
myself
I'm an AI Engineer and Backend Developer with a research interest in Artificial Intelligence and Machine Learning. I build systems that work, from agent infrastructure to production ML pipelines.
My work spans research and engineering across AI, from exploring new ideas and building intelligent systems to taking them into production. I believe intelligence should be universal—not limited by language, geography, or access to resources.
When I'm not building or researching, I run training bootcamps and mentor people, because the best way to learn is to teach.
Publications & interests
Evaluating Parameter-Efficient Fine-Tuning for Cross-Lingual Natural Language Inference: A Case Study on Nepali XNLI
Shapkota, A. & Ghimire, S.
Study of Cross-lingual Natural Language Inference (XNLI) for Nepali using XLM-RoBERTa base, comparing full fine-tuning against parameter-efficient Low-Rank Adaptation (LoRA)
Research Interests
Education
Bachelor of Engineering in Information Technology
Nepal College of Information Technology — Pokhara University
Artificial Intelligence, Data Science and Analytics, Probability and Statistics, Data Structures & Algorithms, Database Management Systems, Operating Systems, Computer Networks, Software Engineering, Calculus & Applied Mathematics, IT System Security
Higher Secondary Education (+2), Computer Science (Science Stream)
Xavier International Secondary School
Computer Science, Mathematics, Physics, Chemistry, English, Nepali
Secondary Education Examination (SEE)
Arunima Secondary School
Certifications
What I work with
Languages
ML & NLP
LLM Platforms
Backend & Systems
Databases & Cloud
Research & Data
Spoken Languages
Where I've worked
Lead Software Engineer (AI & Backend)
@ Kaylos SolutionsMelbourne · Remote
Building multi-modal AI systems, chatbots, voice agents, and intelligent booking workflows using Gemini API, RAG pipelines, and vector databases for semantic search and retrieval-augmented generation.
Python and AI Engineer
@ Omway TechnologiesKathmandu
Built scalable web apps and REST APIs using Django, DRF, and cloud platforms. Developed and deployed AI applications, integrating them into production with best practices. Ensured performance and security through database optimization, caching, CI/CD pipelines, and automated monitoring.
Full Stack Developer
@ National Disability SolutionsSydney · Remote
Expanded an end-to-end disability services platform serving clients and providers. Redesigned core user flows with UX/UI team leading ~40% increase in user engagement. Reduced critical platform issues by ~30% through ongoing technical support and optimization.
Full Stack Developer
@ Neon MarketplaceKathmandu
Supported 10,000+ products and 2,000+ concurrent users with 95%+ frontend performance. Reduced order processing time by 60% and checkout drop-off by 25% via payment gateway integrations. Enabled 50% faster product management for merchants, saving ~10 hours/week.
Selected work
auth-utils
Open-Source npm PackageBuilt and published a zero-dependency, framework-agnostic TypeScript authentication library. Implemented multi-device session management with type-safe adapters for Redis, PostgreSQL, and in-memory storage, alongside CSRF protection and advanced rate limiting. Built RFC 6265-compliant, HMAC-signed, AES-256-GCM encrypted cookie handling with 557+ tests and 95%+ coverage across a 42KB tree-shakeable ESM bundle.
Tapyze
Tap-to-pay Payment System with Digital Wallet & Payment PlatformDesigned the full backend architecture handling transaction processing, balance management, and multi-user authentication. Engineered NFC-based peer-to-merchant payments using ESP32-C3 and PN532, bridging hardware and REST API communication in real time. Architected for horizontal scalability with modular support for future features.
TapEz
Smart Tap-to-Pay SystemBuilt a contactless tap-to-pay system integrated with a Django backend for fast, checkout-free retail transactions. Developed an AI-powered product recognition module using OpenCV, TensorFlow, and a custom-trained scikit-learn model. Integrated a RAG component to power contextual product information and customer assistance during checkout.
Let's build
something together
I'm always open to new opportunities, research collaborations, or a conversation about AI.
© 2026 Amir Shapkota