Joma
Software Engineer
| Phone: | +65 89414036 |
| Email: | axuplus@gmail.com |
| Website: | joma.eu.org |
| GitHub: | github.com/nodfans |
| LinkedIn: | in/ma-xu |
Dedicated software developer with 4+ years of experience developing complex software solutions for private sector clients. Specialised in designing performant, user-friendly applications that streamline operations and boost productivity.
Strong background in database engineering, full-stack web development, and system architecture, with expertise in Java, Python, Linux, AI, and cloud infrastructure. Committed to delivering innovative, scalable applications that address real business challenges.
Key Skills
- Programming & Scripting
- Database Management
- CI/CD
- RESTful API
- Microservices
- Problem Solving
- Self-Motivation
- Analytical Thinking
- Verbal Communication
- Agile & Scrum
Working Experience
Queue Manager
Scale AI (May-2025 - Present)
- Managed contributor experience and retention by providing timely support, resolving issues, and Organising webinars, while fostering a collaborative community through facilitating and moderating discussions across SFT and RLHF projects.
- Served as a liaison between contributors, Single Threaded Owners (STOs), and project pipeline admins to streamline communication and ensure high-quality project alignment across initiatives for timely and effective client delivery.
LLM Training Specialist
Scale AI (Apr-2024 - May-2025)
- Reviewed coding capabilities across diverse programming languages for AI models, focusing on prompt generation, response evaluation, rubric creation, and correction of code snippets, with in-depth analysis of coding tasks.
- Assessed AI-generated responses for instruction following, groundedness, completeness, truthfulness, clarity, verbosity, sentence structure, grammatical accuracy, contextual appropriateness, Markdown rendering, and overall coherence.
- Conducted comparative analysis and scoring of AI-generated responses to assess their quality side by side, by providing detailed written justification to identify claims or factors in the responses that could negatively impact quality.
Backend Engineer
QuanYuFeng Technology (May 2021 – Aug 2021)
- Refactored Java API interfaces and redesigned business logic to improve code maintainability and align with evolving project requirements.
Backend Engineer
AnXiao Technology (Jul 2020 - Jan 2021)
- Completed assigned database tasks including identifying performance bottlenecks, optimising complex SQL queries, and conducting SQL table index invalidation maintenance, resulting in a 21.5% reduction in response time across API calls.
- Integrated third-party APIs, Utilised Docker to wrap 42 microservices as Docker containers, deployed them on local VMs for reliable execution, employed Spark and Hadoop for server log storage, and drafted deployment documentation.
- Contributed to multi-project requirements analysis, engaged in discussions with teammates on the design of 94 PostgreSQL tables, and involved in architecture and Java API development in accordance with design requirements.
Backend Engineer
Fanyu Technology (Nov 2019 - Jun 2020)
- Developed and maintained Java APIs across 8 projects with a microservices architecture, optimised MySQL/Elasticsearch schemas to improve query performance, and deployed 25+ microservices on CentOS using Shell scripts with automated logging and service error monitoring to ensure high system availability.
- Collaborated with frontend teams to integrate APIs with 4 web applications, managed deployment of all services across 3 production environments, configured domain settings for 5+ subdomains, and implemented SSL certificates ensuring secure communications for the entire platform.
Projects
GPT For PDFs | Python, React, OpenAI API, Faiss, Redis, MongoDB
- Implemented a website resembling ChatGPT for PDFs using the open-source LLM framework Langchain, enabling text information capture, extraction, and user Q&A capabilities.
- Integrated Faiss (Facebook AI Similarity Search) to improve prompt response quality via relevant text embeddings and vector similarity scoring, enhancing PDF document analysis.
- Built a responsive UI with React and Tailwind; efficiently leveraged MongoDB and Redis to persist user conversation history, ensuring smooth, seamless interaction and intuitive user experience.
UN IoT Solution | Python, Azure
- Simulated an IoT solution to collect global temperature data addressing climate change in alignment with UN concerns.
- Leveraged Azure IoT Hub to monitor 400+ temperature devices worldwide, ensuring connectivity, raising alerts, and storing large volumes of telemetry data.
- Built a Python service for anomaly detection, triggering alerts when temperature thresholds were exceeded.
Education
Bachelor of Science in Computing Science (First Class Honours), 2024
Coventry University, UK