Haoting Shen

I’m a full-stack software engineer with experience building backend systems, web applications, and cloud-based services. I’ve worked with Python, FastAPI, React, and TypeScript to develop and deploy production-ready applications, and I’m comfortable using AWS services like EC2, S3, and RDS along with CI/CD pipelines.

I’ve built and scaled a full-stack platform that serves over 50,000 users, where I focused on designing reliable data pipelines, improving performance, and delivering a smooth user experience. I enjoy working across the stack and taking products from idea to deployment.

I’m a fast learner and tend to pick up new technologies quickly, especially when working in unfamiliar areas. I’m also detail-oriented and focused, and I like taking ownership of problems and seeing them through to completion in a practical, results-driven way.

Founder & Full-Stack Developer at Roco Kingdom Team Builder (2025-06 – Present)

Independently built and launched a production web platform used by 50,000+ game players to build and analyze optimal PvP team.

• Independently built and launched a production web platform https://rkteambuilder.com used by 50,000+ game players to build and analyze optimal PvP team, using React, TypeScript, TanStack Query, and Tailwind CSS on the frontend with FastAPI, SQLAlchemy, PostgreSQL, and Redis on the backend; containerized with Docker and served via AWS EC2, CloudFront CDN, and RDS.
• Designed 13-model relational schemas with Alembic migrations and 12 idempotent ETL pipelines ingesting Excel data; implemented battle analysis engine for stat formulas, type coverage, and team composition recommendations.
• Implemented concurrent async batching across 6 per-monster Gemini/DeepSeek LLM calls, cutting latency by 75%; layered Redis distributed caching with stampede protection and reference-resolved prompt engineering for accuracy.
• Applied full SDLC with GitHub Actions CI/CD using AWS OIDC, running Pytest on containerized PostgreSQL and Redis, building and pushing Docker images to ECR, deploying via SSM to EC2, syncing frontend to S3, and invalidating CloudFront cache; secured with 5-tier JWT auth, httpOnly cookies, bcrypt, and 4-layer rate limiting.

AI Software Engineer Intern at Eth Tech (2025-09 – 2025-12)

Developed a RAG-based AI knowledge assistant using Python and FastAPI, integrating LangChain and OpenAI APIs for semantic search and answer generation.

• Developed a RAG-based AI knowledge assistant using Python and FastAPI, integrating LangChain and OpenAI APIs for semantic search and answer generation across 500+ internal documents.
• Designed a document ingestion pipeline, recursively chunking 500+ files into 10,000+ segments with 64-token overlap, enabling sub-second top-5 semantic retrieval via Chroma/FAISS indexing.
• Implemented Pydantic schema validation for LLM outputs and created a regex-based block extractor for malformed responses, reducing pipeline failure rate by ~60%.
• Built a two-layer prompt injection guardrail using regex for fast rejection and secondary LLM self-checks, mitigating risks from 30+ attack templates and ambiguous inputs.
• Established a 50+ golden Q&A evaluation harness scored via binary human judgment and integrated regression tests into a GitHub Actions CI pipeline, maintaining over 80% answer relevance despite model and prompt changes.
• Containerized the system with Docker and deployed as a serverless service on AWS ECS Fargate with S3 for document storage and CloudWatch for latency/error monitoring under IAM least-privilege policies.

Frontend Engineer Intern at Bolland International (2024-06 – 2024-08)

Designed and developed the official company website end-to-end using React.js and TypeScript, implementing responsive UI and SEO best practices.

• Designed and developed the official company website end-to-end using React.js and TypeScript, gathering business requirements, implementing responsive UI, and applying SEO best practices to boost organic search traffic by 35%.
• Implemented a dynamic image carousel, page transitions, and lazy-loaded product visuals, reducing 40% load time and driving over 60% user engagement improvement.
• Built a searchable, filterable catalog for 50+ food products, improving customer navigation and product accessibility.
• Engineered dual deployment pipelines for FTP production and GitHub Pages previews using custom Node.js scripts.
• Integrated Google Analytics to track SEO performance, improving bounce rate by 15% through UX optimizations.

Software Engineer Intern at FrogHire.ai (2023-05 – 2023-08)

Spearheaded backend development in Python of FrogHire.ai, a web browser extension that streamlined job searches across multiple platforms.

• Spearheaded backend development in Python of FrogHire.ai, a web browser extension that streamlined job searches across LinkedIn, Handshake, Indeed, Glassdoor, and Google jobs, achieving 40,000+ active users in its first year.
• Collected post-launch user feedback and iteratively refined backend models to improve result relevance.
• Designed and implemented NLP pipelines with spaCy to analyze job descriptions, using tokenization, POS tagging, and Named Entity Recognition to improve structured extraction of critical job requirements and qualifications.
• Built NLP extractors for degree requirements using Python RegEx with 98% accuracy and for work experience using spaCy Matcher and Fuzzy Search with 92% accuracy. Implemented sponsorship detection with custom rule-based dictionaries and negation handling to classify whether job postings require U.S. citizenship or offer visa sponsorship.
• Developed Pytest test suites to simulate real-world and edge cases, boosting result accuracy from 60% to 95%.

B.S. in Computer Science, Minor in Mathematics – Washington University in St. Louis (2021-09 – 2025-05)