Team
ABConvert-TeamB
This team is at maximum capacity.
Project Concept
No description has been added yet.
Entry
Status: Submitted
Last saved: December 11 at 9:34 PM CST
Team Roster (team is at max capacity)
Message board not available for this team yet.
Jeffrey Liu Team Lead RSVP Approved
Head of Engineering at ABConvert
Project Architecture & Mobile App Development Led the overall system architecture design and built the Flutter mobile client. Responsible for implementing the real-time WebSocket connection to handle the bi-directional data flow. Integrated OpenAI Whisper for voice capture and optimized the mobile camera feed to send efficient image frames to GPT-4V for analysis.
Jeffrey Liu is the Head of Product, Personalization Cloud at Appier, and also holds the title of Head of Engineering at ABConvert. With over 20 years of experience, he is based in Taipei City, Taiwan. Jeffrey studied Information Management at National Taiwan University. His current employment status indicates he is self-employed, seeking co-founders, and seeking investors, while also taking on a Tinkerer role as a Technical Product Manager.
Seeking co-founders, seeking investors, predictive/generative AI, personalization & recommendation systems, ML infrastructure, conversational commerce
Jacklyn Cheng RSVP Approved
AI Engineer at ABConvert
Next.js Web App Development Focused on the web client interface using Next.js. Developed the responsive cooking dashboard UI and handled the frontend integration of ElevenLabs' Text-to-Speech API, ensuring the agent's voice feedback was synchronized with visual cues in the browser environment.
Data Scientist with 3+ years of experience in full-cycle data projects, from analysis to deployment.
Proficient in Python, SQL, and ML/DL frameworks including Scikit-learn, LightGBM, and PyTorch.
Experienced in building RAG pipelines to integrate generative AI into analytics workflows.
I have experience with PyTorch and Scikit-learn and am currently expanding into LLM orchestration. I’m looking to connect with others experimenting with advanced RAG patterns (e.g., hybrid search, re-ranking) to enhance business decision support and optimize complex operational workflows.
I recently completed an Automated Schema Documentation solution to address data governance challenges.
I designed a specialized RAG pipeline that ingested ETL scripts and pipeline lineage to automate schema creation. The project featured a three-stage LLM workflow (Code Retrieval → Feature Extraction → Schema Writing), where I utilized Chain-of-Thought prompting and strict output constraints to ensure consistent documentation from raw code.
Yuan Jung Hsieh RSVP Approved
Customer Success Manager at ABConvert
Project Architecture & Use Case Design Defined the core agent workflow and designed the "French Omelette" user journey. Engineered the specific system prompts for GPT-4V and GPT-4 to ensure the agent could accurately interpret visual cooking stages (like egg texture or butter browning) and provide safe, context-aware instructions.
Customer success manager at ABConvert
how to build a product from scratch (idea, development, testing....)
no
MING HUNG TSAI RSVP Approved
Backend Engineer at ABConvert
Full-Stack Web App Development Managed the full-stack integration connecting the Next.js frontend with the backend services. Orchestrated the API logic for OpenAI GPT-4, managing the state of the cooking session and handling the asynchronous data exchange between the user’s inputs and the agent’s reasoning engine.
Ming-Hung Tsai is a backend engineer based in Taipei, Taiwan, specializing in high-throughput distributed systems and cloud-native infrastructure. He has built and scaled payment and order-management services across multiple countries at foodpanda, improving latency, reliability, and cost efficiency through pragmatic architecture and performance tuning.
Recently, he developed “Ticket Master,” a Kafka-based ticket reservation system reaching 83,000 reservations per second, and continues to contribute to a hiking safety support platform that delivers real-time trip updates via a serverless Go and GCP stack. With strengths in Go/Java, Kubernetes, and observability, Ming-Hung focuses on turning complex systems into fast, resilient, and measurable products.
Building and scaling distributed, event-driven backend systems (especially Kafka/Dataflow architectures) and pushing performance/reliability limits through rigorous load testing and data-driven optimization.
Location sharing and communication apps for hiking safety