Yucheng (Cecilia) Zhang

AI / ML Research Engineer

Yucheng (Cecilia) Zhang

About

About me.

I'm Yucheng (Cecilia) Zhang, an M.S. student in Computer Science at Northwestern University. Before that, I earned my B.S. in Mathematics with a Computer Science minor from The University of Texas at Austin.

The world is full of messy data, scattered information, and more choices than anyone can weigh. Studying math gave me something I'm still grateful for: the ability to sit quietly with a problem and think it through on my own. I'm fascinated by the logic hiding behind things, and I want to use applied methods to analyze it, explain it, and maybe create something with real meaning and value. Honestly, that process is where I find a lot of my joy.

I try to remember that behind most data there are people: someone waiting for a diagnosis, a student stuck on a bug late at night, a family hoping a loan goes through. I want to work carefully and honestly, stay curious about what I don't understand yet, and be kind to the people I work with. I'm still learning, and I hope I always will be.

Current focus: LLM systems, applied ML, and model evaluation for real users.

Technical base: Python, SQL, React, Flask, PostgreSQL, Docker, Scikit-learn, TensorFlow, PyTorch.

Also: photography, painting, and writing for fun on Medium.

Work

Engineering and industry experience.

Data Scientist Assistant Intern - Image Restoration

Chinese Academy of Sciences | Beijing, China | Jul. 2023 - Sep. 2023

Deployed image-restoration modules on Alibaba Cloud, implementing VAE loss, Gaussian filters, LSGAN components, a TensorFlow CNN, and GFPGAN restoration workflows.

Financial Data Development Engineer - Fraud Detection

Caitong Securities Ltd | Hangzhou, China | Jul. 2024 - Aug. 2024

SQL ETL over 10M+ daily transactions with behavioral fraud features (transfer timing, test-then-transfer patterns, account dormancy), and a weighted risk-scoring model tuned by backtesting that cut false positives by 2%.

Data Science & AI Engineering Intern - Customer Analytics

Joblogic-X Corporation | Dallas, TX | Nov. 2023 - Jun. 2024

Forecast daily demand for 19+ dishes with a Prophet + XGBoost pipeline tuned with Optuna (R² up to 0.98, MAPE as low as 7.4% on held-out data), and scored social-media influencers from 500+ posts, finding image posts drove 5.6x more engagement than video.

Research

Research experience and publications.

Publications

  1. Jin, Z., Zhang, Y., Islam, R., & Ahuja, K. (2026). Synchrony is not readability: Evaluating lip-sync metrics on translated visual speech in English and Mandarin. ACL ARR (under review). OpenReview
  2. Miao, H., Wang, Z., Zhang, Y., ... & Tong, Y. (2026). New kiwifruit quality grading methods based on image multi-feature fusion and ResTNet model. Expert Systems with Applications, 297, 129507. DOI
  3. Xu, M., Zhang, Y., Zhang, Y., Liu, X., & Qing, K. (2024). EEG biomarkers analysis in different cognitive impairment after stroke: An exploration study. Frontiers in Neurology, 15, 1358167. Link
  4. Zhang, Y., Zhang, Y., Jiang, Z., Xu, M., & Qing, K. (2023). The effect of EEG and fNIRS in the digital assessment and digital therapy of Alzheimer's disease: A systematic review. Frontiers in Neuroscience, 17, 1269359. Link
  5. Zhang, Y. (2023). EEG-Based on Emotion Recognition Using Machine Learning. In IEEE ICIPCA (pp. 226-232). IEEE
  6. Li, D., Zhang, Z., & Zhang, Y. (2023). Public compassion under the mediation of suffering: Taking the videos of "Second Uncle" on social media as an example. Media and Communication Research, 4(5), 46-58. Link
Creative Work

Photography and painting.

The same attention that goes into a modeling pipeline (noticing pattern, structure, and what's actually there versus what I expect to see) shows up here too, just slower and without a deadline.

And sometimes I just write, about whatever I've been thinking about lately. You can find those on Medium.

Contact

Say hi. I'm happy to talk about anything.

I love meeting people who see the world differently from me. Every conversation leaves me with something new to think about. So feel free to reach out, whether it's about research, work, painting, or something completely random.

ceciliayczhang@gmail.com