AI Engineer building speech recognition, document QA, and career-guidance systems — from research prototype to deployed AWS backend.
Computer engineer with a research background in machine learning and deep learning, currently working as an AI engineer on production systems.
I like taking a model from a research idea to something that actually runs in front of users — training it, wiring it into a backend, and deploying it. Past work spans Nepali speech recognition, LLM-based document QA, and image tagging systems, alongside the AWS infrastructure that keeps them running.
Comfortable moving between research and infrastructure: the same week might involve fine-tuning a transformer and shipping the CI/CD pipeline that serves it.
Research prototypes and shipped systems, from speech models to production AI backends.
A career support platform offering an interactive roadmap display and AI-powered conversational counseling, built on the OpenAI API with an AWS backend.
OpenAI API · AWS · FastAPIAn intelligent chatbot on Llama 2 and Vicuna that answers user queries from documents and a SQLite database.
Llama 2 · VicunaA CNN–Transformer speech recognition system for Nepali. Published in LDK 2023.
TensorFlow · KerasAutonomously learns image tags from internet data, building a tag hierarchy through embeddings.
EmbeddingsAn image captioning system built on a pretrained CNN with a transformer decoder.
CNN · TransformerA website search feature offering smart suggestions based on a site's own content.
NLPAn online platform for listing and booking rental homestays, hotels, and houses.
Full-stackAn Android app that tracks a phone and syncs messages and call logs to the web in real time.
Android · KotlinPaudel et al., "Large Vocabulary Continuous Speech Recognition for Nepali Language using CNN and Transformer," LDK 2023.
Based in Pokhara, Nepal. Open to conversations about AI engineering, applied research, and interesting backend problems.