About me

I’m Amit, a Full Stack Software Engineer with 5+ years of experience delivering scalable platforms and production-grade AI systems. I’ve improved system reliability by 40% and supported platforms serving 1,000+ stores. I’ve built and deployed RAG pipelines, agentic AI systems, and distributed inference infrastructure across Kubernetes and HPC environments.

My expertise spans full-stack architecture, CI/CD and inference optimization for large-scale ML systems. I’m a published AI researcher with an award-winning paper (3 citations), and I’m currently pursuing an MS in Computer Science (AI) at the University of South Dakota. I’m a Toptal Top 3% elite developer, having delivered production AI systems for clients across legal, healthcare, and compliance industries.

Highlights

  • Web development icon

    Software Development

    Building robust and scalable backend and full-stack systems.

  • design icon

    AI & ML Development

    Designing intelligent systems that analyze data and automate insights.

  • Web development icon

    Web & Data Systems

    Delivering web applications, dashboards, and data-driven solutions

  • mobile app icon

    Mobile Apps

    React Native apps, visualization tools, end-to-end mobile solutions

Resume

Education

  1. University of South Dakota

    August 2024 — Present  |  Vermillion, South Dakota, USA

    MS in Computer Science - Artificial Intelligence

  2. IOE, Pulchowk Campus, Tribhuvan University

    November 2014 — September 2018  |  Pulchowk, Lalitpur, Nepal

    Bachelor’s Degree in Computer Engineering

Experience

  1. Research Assistant

    Spring 2026 – Present  |  University of South Dakota, Vermillion, SD

    • Conducted research on Reinforcement Learning with LLMs and Agentic AI Frameworks; built and benchmarked baselines including Chain-of-Thought, ReAct, Reflexion, and ExPeL — developing enhanced variants that improved reasoning stability and task performance by up to 30%.
    • Engineered end-to-end experimentation pipelines (training, evaluation, deployment) on the Nautilus HPC cluster using vLLM, OpenAI SDK, Kubernetes pods, and Dockerized environments.
    • Led VideoAgent optimization project — iteratively improved baseline agent architectures and achieved measurable gains in multi-step video reasoning performance.
    • Investigated CNN robustness under distribution shifts; resulted in one publication with 3 citations (3rd Best Paper Award). A second research line is under active experimentation.

  2. Senior Software Engineer II

    Oct 2023 – Aug 2024  |  LIS Nepal, Lalitpur, Nepal

    • Inteliome AI Chatbot: Built a full-stack application enabling file uploads and conversational Q&A over document contents. Used Next.js with Jotai for global state and React Query for async data fetching and caching. Developed a highly responsive, lightning-fast UI with Tailwind CSS, Shadcn, and Radix UI.
    • Architected the RAG-based backend with multi-source ingestion, async processing via Celery and RabbitMQ, and OpenAI + LangChain pipelines — improving system reliability by 40% and significantly reducing production incidents.
    • Mentored 3 engineers and standardized reusable backend AI service patterns, reducing onboarding time by 25%.

  3. Software Developer (Freelance) — Toptal Top 3%

    Feb 2022 – Aug 2024  |  Remote

    • Hair Density Tracker (Sunai): Bootstrapped a React app for tracking scalp hair density over time using high-resolution image analysis. Implemented Firebase authentication, Redux + Context API for state management, React Query for async API calls, and digital record tracking of hair counts. (React, Redux, Firebase, React Query, Material UI, SCSS, AWS)
    • LegalFit CRM: Developed a premier web marketing platform for US law firms. Built multiple responsive layouts and blocks for custom website creation, following Core Web Vitals and Lighthouse guidelines. Authored Jinja macros for templating across the platform. (Jinja, SCSS, JavaScript, Python, Material UI, AWS, Bootstrap)
    • Compliance Q&A Backend: Built an AI-powered compliance system handling 1,000+ automated regulatory queries using agentic RAG architecture with LangGraph, vector DBs, Kafka event ingestion, and MongoDB. (LangGraph, VectorDB, MongoDB, Kafka, Node.js, Python)

  4. Senior Software Engineer

    Aug 2021 – Feb 2022  |  LIS Nepal, Lalitpur, Nepal

    • Gocchii Compliance Tools: Built compliance management software for tracking gift exchanges and transactions via Plaid, and submitting/reviewing marketing materials for compliance officers across 5+ office locations. Served managers and employees with role-based access — reducing manual regulatory effort by 50%. (React, Django, Redux, MUI, Plaid, PostgreSQL, JWT)
    • Project Management & Timesheet System: Built an internal tool for 200+ employees covering project tracking, resource management, sprints, timesheets, and weekly reporting — improving reporting accuracy by 35%. Added REST API endpoints to support the React frontend. (React, Laravel, PostgreSQL, MySQL, REST APIs)

  5. Software Engineer

    Aug 2020 – Aug 2021  |  LIS Nepal, Lalitpur, Nepal

    • Curbside Pickup & Contactless Retail: Developed back office, order interface, location service, and pickup scheduling for a retailer with 1,000+ locations across 10+ time zones. Incorporated Google Maps API for geolocation, time-zone-critical logic, and data insights. (Node.js, MySQL, Sequelize, Express.js, Google Maps API, TypeScript, Docker, Jenkins CI/CD)
    • Nepal Stock Exchange (NEPSE): Integrated a local payment gateway into the Trade Management System and implemented an auction engine for matching buy-and-sell orders during market sessions. Achieved 90% test coverage. (Java, Angular, Scala, Redis, RabbitMQ, JUnit, Mockito, Jenkins)
    • Robling ETL Monitoring App: Built a React Native mobile app for monitoring Snowflake ETL batch processes with push notifications via Firebase. REST API built with Python Flask — reducing system downtime by 20%. (React Native, Python Flask, Firebase, Snowflake)

  6. Associate Software Engineer

    May 2019 – Aug 2020  |  LIS Nepal, Lalitpur, Nepal

    • OpinionShield Android App: Developed an Android app for rating businesses, where candidates were ranked based on linked account transactions retrieved via the Plaid API. Published multiple versions to the Play Store; grew to 5,000+ active users with in-app purchase support. (Java, Android, RxJava, Retrofit, Plaid, In-App Purchases)
    • Oracle Xstore POS: Configured and integrated Oracle Xstore point-of-sale systems for retail clients using the Oracle POS system as a base, streamlining Java-based workflows and reducing deployment setup time by 30%.

My skills

  • Python
    95%
  • JavaScript / TypeScript
    95%
  • Node.js
    95%
  • Java
    85%
  • React / Next.js
    95%
  • Django / FastAPI / Flask
    93%
  • React Native / Android
    80%
  • AI / Machine Learning (LLMs, CNNs, RAG)
    88%
  • PyTorch / TensorFlow / scikit-learn
    85%
  • PostgreSQL / MySQL / MongoDB / Redis
    90%
  • Docker / Kubernetes / CI-CD
    75%
  • AWS / Firebase
    75%
  • REST / GraphQL / Git / Agile
    90%
  • Kafka / RabbitMQ
    85%

Portfolio

Contact

Contact Form