Hi, I'm Ashar UL Huda
Software Engineering student building AI-powered full-stack applications. I focus on AI engineering, backend systems, and user-facing products, with recent work across chat workflows, FastAPI services, MongoDB schemas, and React + TypeScript interfaces.
Irvine, CA
Open to SWE, AI, backend, and full-stack roles. Open to relocation.
About me
Software engineering, AI systems, and full-stack products.
I'm a Software Engineering and Informatics student at the University of California, Irvine, specializing in Human-Computer Interaction and graduating in June 2026.
I'm interested in AI systems, backend architecture, and full-stack products that turn complex workflows into usable tools. My recent work includes AI chat workflows, API design, database modeling, document upload flows, and React interfaces.
My main capstone project is LoanFit CoPilot, an AI-powered loan assistant for loan officers. I'm building the React + TypeScript chat UI, session persistence, MongoDB data models, and a FastAPI backend connected to an AI pipeline for loan application analysis.
Quick bits about me:
- Based in Irvine, CA
- SWE, AI, backend, and full-stack focus
- Open to relocation
- GitHub: ahuda07
I'm always happy to connect about software engineering, AI engineering, backend, and full-stack opportunities.
Skills
Organized around the work I want to do: AI systems, backend architecture, and full-stack products.
Languages
Core programming languages for backend, systems, and web.
Full-Stack
Frontend, API, and product-building frameworks.
AI & Data
ML, NLP, search, and AI application tooling.
NumPy
PyTorch
OpenCV
BeautifulSoup
NLTK
Gradio
MCP
Experience
Recent engineering experience:
Software Developer Intern
Easy Street Capital
Jun 2026 - Sep 2026
Backend Software Engineer - Senior Capstone
Champions Funding
- Built FastAPI backend services for LoanFit CoPilot, an AI loan-assistant platform that screens borrower scenarios against loan-product requirements such as FICO, occupancy, documentation type, DSCR, and LTV.
- Developed backend logic for a two-agent workflow, passing structured borrower and loan context from a Copilot agent to a Coach agent that identifies qualification gaps and suggests improvement paths.
- Modeled MongoDB collections for borrower sessions, agent message history, and uploaded loan documents to support persistent multi-turn conversations.
- Defined REST API contracts and JSON payloads that connected backend agent responses with React + TypeScript chat sessions and document data.
Jan 2026 - Jun 2026
Resume
A quick CV snapshot for software engineering, AI engineering, backend, and full-stack roles.
Education
University of California, Irvine
B.S. in Software Engineering and Informatics
Specialization in Human-Computer Interaction
Expected June 2026
Target Roles
Software Engineering, AI Engineering, Backend Engineering, and Full-Stack Engineering. Based in Irvine, CA and open to relocation.
Engineering Snapshot
- FastAPI backend services for LoanFit CoPilot, including loan-product screening logic and two-agent AI workflows.
- MongoDB data modeling for borrower sessions, agent message history, and uploaded loan documents.
- Search infrastructure with post-index processing, tf-idf, PageRank, SimHash near-duplicate detection, and Streamlit querying.
- Multi-agent AI code-generation workflows with LangGraph, FastMCP, Google Gemini 2.5, and Gradio.
Projects
Projects spanning AI workflows, backend systems, and full-stack products:
Welcome to LoanFit Copilot!
Sign in to LoanFit Copilot
Welcome back! Please sign in to continue
LoanFit CoPilot
- Built FastAPI backend services for an AI loan-assistant platform that screens borrower scenarios against loan-product requirements such as FICO, occupancy, documentation type, DSCR, and LTV.
- Developed backend logic for a two-agent workflow, passing structured borrower and loan context from a Copilot agent to a Coach agent that identifies qualification gaps and suggests improvement paths.
- Modeled MongoDB collections for borrower sessions, agent message history, and uploaded loan documents to support persistent multi-turn conversations.
- Defined REST API contracts and JSON payloads that connected backend agent responses with React + TypeScript chat sessions and document data.
React
TypeScript
FastAPI
Python
MongoDB
AI/LLM pipeline
Pirate Search Engine
- Engineered Python post-index processing routines for a large ICS web corpus of roughly 55,000 documents, scanning a byte-offset inverted index to generate bookkeeping artifacts loaded by the query engine at startup.
- Improved query-time performance by precomputing tf-idf and PageRank offline and applying SimHash-based near-duplicate detection, moving expensive computation off the query path.
- Built search pipeline components across corpus processing, indexing, ranking, and the Streamlit query interface, with a focus on post-index data generation and ranking artifacts.
Python
NLTK
Streamlit
Inverted Index
tf-idf
PageRank
SimHash
Trailblazer - Biking Trail Finder
- Built the data layer for a biking-trail finder web app during IrvineHacks Winter 2024, integrating RapidAPI TrailAPI with zip-code geocoding and difficulty-filtering logic on API responses.
- Implemented MongoDB Atlas Data API CRUD operations for user records and saved-trail favorites, avoiding a custom Express backend and reducing hackathon development overhead.
- Connected API results to React frontend components, rendering filtered trails in a Bootstrap/Ant Design carousel with Google Maps location markers.
React
JavaScript
MongoDB Atlas Data API
RapidAPI
Google Maps API
axios
Bootstrap
SportsPlayerScout
- Developed a 3-agent AI code-generation system where Analyzer, Developer, and Tester agents translate natural-language requirements into runnable Python code and test cases.
- Integrated Google Gemini 2.5 Flash with a LangGraph + FastMCP workflow, enabling iterative code generation, review, and testing loops across agents.
- Implemented the Gradio interface for prompt intake, generated-code display, end-to-end execution flow, and reproducible setup documentation.
Python
LangGraph
MCP/FastMCP
Google Gemini 2.5
Gradio
Get in touch
Reach out about software engineering, AI engineering, backend, or full-stack opportunities.
hudaau@uci.edu
You can also find me on GitHub and LinkedIn.