Sushant Karki

I started my professional career as a full-stack web developer at Minma, Inc., a startup in Japan, where I spent eight years building and maintaining an online marketplace. Over time, I developed an interest in Machine Learning and transitioned into this field in 2022 by enrolling in a graduate program at the University of Maryland, College Park. Now, as a Machine Learning Engineer, I work on exciting projects that combine software engineering and AI to solve real-world problems.

What I’ve Worked On:

• Full Stack Web Development:
  Designed and implemented high-availability web applications, streamlining deployments and improving system reliability.
  • Curama: A platform connecting users with service providers for home and business needs.
  • Govskills.io: Developed tools to simplify federal job searches and enhance user accessibility.
• End-to-end RAG Application Development:
  Developed Retrieval-Augmented Generation (RAG) systems, improving text analysis precision and relevance for RegTech applications.
• Research Projects:
  • Text Style Transfer with LLMs.
  • Automated unit test generation for C codebases using open-source LLMs.
  • Disentanglement of style and content in speech.

I’m currently interested in exploring how the brain learns, adapts, and retains knowledge efficiently. By studying these processes, I aim to apply neuroscience-inspired principles to develop AI systems that are more adaptable, resilient, and capable of continuous learning.

Feel free to browse my portfolio to learn more about my work!