台北市北投區經歷不拘碩士以上
✅Qualifications
▪️Holds a Master’s degree or has equivalent hands-on experience.
▪️Has experience with machine learning frameworks(e.g., PyTorch, HuggingFace, PyTorch Lightning).
▪️Is comfortable coding in Python, and familiar with SQL, Git, and Web API tools like Flask or FastAPI.
▪️Has experience working on machine learning projects — from experimentation to deployment.
▪️Enjoys working in a team and communicates ideas clearly and effectively.
If you’re curious, hands-on, and always up for a new challenge — that’s a big plus!
▶About the Job◀
Hi! We’re the DCS Team(Department of Cognition and Sensing)— a vibrant group of machine learning engineers, data scientists, and software architects who love exploring new ideas and building meaningful solutions with AI.
We work on cutting-edge machine learning applications, especially in the areas of Computer Vision(CV)and Multi-modal AI. Whether it’s extracting key data from documents, generating image captions, or building smart visual search systems — we’re passionate about bringing intelligence into daily workflows and helping people save time and energy.
Our team culture is open, experimental, and supportive. We’re not afraid to try new things, break stuff, and celebrate wins — big or small. You'll get the chance to work on real-world problems, collaborate across teams, and see your ideas come to life in products used by real customers.
▶Our Focus Areas◀
We work on:
▪️Visual Document Understanding(VDU)
Extracting meaningful data from complex documents like receipts, invoices, and contracts.
▪️Key Information Extraction(KIE)
Pulling out key details like names, dates, and prices from scanned or digital documents.
▪️Image Captioning
Teaching machines to describe what they see — helpful for accessibility, search, and context.
▪️Visual Search(Multi-RAG)
Enabling smart search experiences by combining image and text queries.
▪️Visual Question Answering(VQA)
Answering questions about visual content, combining CV and natural language understanding.
▪️Multi-modal Applications
Building systems that integrate multiple inputs — like image + text — to solve complex tasks.
We collaborate closely with other business teams to explore real-world use cases, deliver solutions, and iterate quickly. You’ll also have the opportunity to attend tech conferences, study SOTA papers, and share what you learn with our internal community.
▶What Your Day-to-Day Might Look Like◀
Life at DCS is a mix of research and engineering
Some days, you’ll feel like a researcher: reading papers, designing experiments, preparing datasets, or testing new model architectures. Other days, you’ll be deep into engineering work — building internal frameworks, packaging models into APIs, and deploying models at scale.
Here’s a brief overview of our daily rhythm:
▪️Small, mission-based squads working on different goals(VDU, visual search, etc.), with lots of cross-team collaboration.
▪️Warm-up meetings to sync up on goals, share progress, and ask for support.
▪️Journal clubs to discuss the latest research, tools, or techniques.
▪️Demos and retros to present your work and reflect on what’s working.
⭐We love building, sharing, and growing together — both technically and personally.
▶Responsibilities◀
As part of the DCS team, you’ll:
▪️Design and build ML-powered solutions — from model training to wrapping them into APIs.
▪️Stay updated on the latest SOTA research and apply novel ideas in real-world scenarios.
▪️Collaborate with other business teams to design ML-driven features that meet customer needs.
▪️Optimize and deploy models to run efficiently across various environments — including on-prem devices and cloud infrastructure.
Ready to join us and build smarter things together?
We’d love to hear from you — whether you’re passionate about machine learning, excited about vision or multi-modal challenges, or simply curious about what we’re building at DCS. Let’s talk :)