Understand people
Reconstruct the user’s context, friction, and motivation before rushing to an answer.
Hi, I’m
Ding Yixiao
AI Product Manager · AIGC · RAG
M.A. candidate at Nanjing University, Class of 2027, with hands-on experience in consumer multimodal generation and enterprise RAG products. I turn complex business scenarios into model capabilities, product decisions, and measurable evaluation systems.
01
AIGC Product
Multimodal generation · Scenario design · Model fit
02
RAG Delivery
Knowledge governance · Chunking · Retrieval
03
Evaluation
Test sets · Metrics · Bad-case analysis
04
Prompt Engineering
Few-shot · Structured prompts · Task instructions
2AI product roles at leading companies
27+Specialized e-commerce image scenarios
300+Industrial knowledge records
IELTS 8.0English · French Level 4
ROLE PHILOSOPHY × SELF-ASSESSMENT
Good products and sustainable growth begin with understanding people.
I see the role as a continuous loop: understand the real need, define the right problem, validate decisions with data, and turn strategy into an experience users can feel.
“Stay sensitive to context, honest about results, and patient with execution.”
Define the problem
Turn an ambiguous request into a clear goal, meaningful metrics, and testable hypotheses.
Drive the outcome
Use product, content, and collaboration to ship, learn from reality, and iterate.
Campus & Global Experience
Making things happen on campus,
learning across cultures.
At Nanjing University and Ocean University of China, I took active roles in student leadership, campus events, and brand partnerships. Learning experiences in the UK, the US, and Switzerland taught me to adapt quickly across languages and cultures.
Nanjing University
M.A. in English Language and Literature
First-Class Scholarship · Graduate Student Union leadership · Organized 10+ major campus events · Secured collaborations with iFLYTEK and SHEIN
Ocean University of China
B.A. in English
National Scholarship · Provincial Outstanding Graduate · Class monitor and Student Union experience · Minor in International Organizations and Global Governance at Peking University
Honors
- 01National Scholarship
- 02Nanjing University First-Class Scholarship
- 03Provincial Outstanding Graduate
International Experience
Learning and living
across cultures.
From a year at a primary school in Manchester to the Penn summer program and UNITAR study in Geneva, these experiences shaped my sensitivity to language, context, and how people understand the world differently.
- UNITAR study program in Geneva, Switzerland
- Innovation and Leadership English Summer Program at the University of Pennsylvania
- One year at a primary school in Manchester, with early immersion in an English-speaking classroom and cross-cultural environment
AI Product Experience
Turning model capability into business outcomes
From consumer multimodal generation to enterprise RAG support.
01
KuaishouProduct Manager · Kling/KOLORS E-commerce Edition
KuaishouBringing image-generation models into the e-commerce visual production workflow across product images, AI try-on, promotional posters, and detail-page assets.
81%+Monthly feature penetration
+66%Asset production efficiency
83%+Useful image rate within 5 turns
Needs & Scenarios
Researched production workflows across fashion, beauty, and 3C merchants; consolidated 20+ representative needs and defined product consistency, industry templates, and batch generation as the core optimization direction.
Model Capability
Contributed to a product-ID recognition, feature extraction, structured prompt, scene generation, and consistency-check workflow; optimized product control, human poses, and background styles for try-on and contextual display.
Product Delivery
Contributed to PRD and interaction design, built a template system covering 8+ categories and 30+ scenarios, and worked with algorithm and engineering teams on 12 evaluation metrics and two model test rounds.
Impact
Completed the MVP iteration in five months and supported 27+ specialized scenarios. Monthly penetration reached 81%+, merchant asset-production efficiency rose 66%, and useful-image rate within five conversational turns reached 83%+.
02
SiemensProduct Manager · AI Xiaoxi Industrial Assistant
SiemensBuilt RAG + AIGC capabilities for 24/7 technical support across PLC, inverter, and servo product scenarios.
300+Technical knowledge records
+30%Knowledge hit rate
+50%Repeat-query efficiency
Knowledge System
Structured manuals, technical documents, FAQs, fault codes, and parameters into a tagged knowledge system; delivered 300+ technical records and 150+ representative fault Q&A samples.
RAG Optimization
Worked on document chunking, cleaning, structuring, and retrieval enhancement for product models, parameter names, and fault codes; constrained response structure and terminology through prompts.
Model Evaluation
Built an industrial Q&A evaluation set and assessed knowledge hits, answer accuracy, diagnosis effectiveness, and parameter matching; analyzed model confusion and citation errors as bad cases.
Product Iteration
Designed model recognition, guided diagnosis, related-knowledge recommendations, and feedback features. Core knowledge hit rate improved by about 30%, while repeat technical-query handling efficiency improved by about 50%.
0–1 Project
Turning a cross-cultural need into a product loop
From user research and architecture to a tested product demo.
Connect · Co-create · Share
Make culture understood
through shared creation.
Interests→Bilingual exchange→Co-creation
BridgeYouth
A Bilingual Co-creation Platform for Chinese and Global Youth
Overview
Designed a bilingual platform around content co-creation, interest-based connection, and cultural experiences to reduce language friction and make cross-cultural exchange more continuous.
My Contribution
Owned user research, prioritization, information architecture, PRD, roadmap, prototype, reviews, acceptance, and testing. Coordinated engineering and UI delivery and drove 20+ experience improvements.
20+Experience improvements
+40%Publishing efficiency
85%+Match completion
Interest Tags →Smart Match →Bilingual Exchange →Co-creation →Community
View Project Details
User Problem
- High language friction and missing cultural context
- Limited mechanisms for sustained content co-creation
- One-off online interactions that rarely become lasting connections
Solution
- Bilingual feed with translation and cultural context
- Interest tags, intelligent matching, and online pairing
- Culture stories and events that connect exchange, creation, and distribution
DIGITAL TWIN · DEEPSEEK
Talk to my digital twin
It answers only from my résumé and project history. Ask as a recruiter, collaborator, or future teammate.
Ding Yixiao · Digital Twin
Answers are generated by DeepSeek and grounded in résumé facts
Beyond Work
What keeps me curious
RedNote · Content Creation
Documenting life while observing how content is understood and shared.
Lifestyle / Content Creation / Everyday Observations
Reading
Long-term interest in fiction, nonfiction, and social and cultural writing.
Writing
I keep notes on everyday observations and write regularly.
Running / City Walks
I like exploring cities on foot and at a slower pace.
Cross-cultural Observation
I stay curious about how language and internet culture differ across countries.