Weijin Zhu
Full Stack AI Engineer
I design and build AI-first web products — agentic interfaces, reliable LLM integrations, and scalable backend systems that put machine learning into production.
About Me
I'm an award-winning data scientist with hands-on experience designing, building, and deploying production-quality machine learning and generative AI solutions. I specialize in RAG-based conversational systems, large-model fine-tuning, NLP analytics, and cloud-native model deployments across Azure, GCP, and AWS.
Recent highlights include delivering a Retrieval-Augmented Generation chatbot in Azure that enabled significant customer support and lead-generation benefits, conducting topic modeling and sentiment analysis on customer recordings, and deploying CNN-based defect detection models. Earlier work includes automation and ETL platform engineering at M&T Bank and research on SonicPrint (ACM MobiSys 2020 Best Paper Award).
I focus on building reliable, scalable, and measurable ML systems and agentic workflows that solve real business problems. I hold the Microsoft Azure AI Engineer Associate (AI-102) certification and an AI Agents fundamentals certification from Hugging Face.
I currently serve as a Senior Full Stack AI Engineer at PwC (08/2025 — Present), where I work on agentic workflows, multi-tenant cloud deployments, and AI-powered solutions for clients.
Technologies I work with:
Where I've Worked
Senior Full Stack AI Engineer @ PwC
08/2025 — PresentDelivered AI-powered solutions to clients by engineering agentic workflows and multi-tenant cloud infrastructure, optimizing agents through context engineering, and piloting cutting-edge developer and automation tools to solve complex business problems.
Data Scientist @ GAF
07/2022 — 03/2025Developed Generative AI-powered chatbots leveraging Retrieval-Augmented Generation (RAG) in Azure cloud, delivering $1.5M annual benefits. Fine-tuned GPT models and performed topic modeling to extract customer insights from phone recordings, earning the Spotlight Award. Led NLP-based sentiment analysis for claims data, improving satisfaction and reducing costs by 20%. Built and deployed CNN models for automated defect detection in manufacturing. Implemented regression models to predict Customer Lifetime Value.
Software Engineer - Automation @ M&T Bank
07/2021 — 07/2022Managed platform compliance and life-cycle dashboards for server management. Automated ETL processes and codebase updates using Ansible, reducing processing time by 30% and error rates by 22%. Developed Flask and Angular web app for compliance reporting, increasing data accessibility and reducing manual report generation time by 46%. Winner of M&T Bank Hackathon 2021 for ML algorithm predicting optimal ATM locations.
Research Assistant @ Center For Identification Technology Research
05/2020 — 11/2020Led pilot study and survey on public concerns in biometrics, revealing fairness and reliability as key issues. Proposed mitigation strategies and performed statistical, qualitative, and quantitative analysis.
Research Assistant @ Embedded Sensing and Computing Lab
04/2019 — 08/2019Developed SonicPrint, a novel identification method using fingerprint-induced sonic effects. Preprocessed sound signal data, designed feature extraction algorithms, and implemented classification models (LR, SVM, RF). Work won ACM MobiSys 2020 Best Paper Award.
Publications
SonicPrint: a generally adoptable and secure fingerprint biometrics in smart devices.
Rathore, Aditya Singh; Weijin Zhu; Afee Daiyan; Chenhan Xu; Kun Wang; Feng Lin; Kui Ren; Wenyao Xu
A Survey on Heart Biometrics
Rathore, Aditya; Li, Zhengxiong; Zhu, Weijin; Jin, Zhanpeng; Xu, Wenyao
ACM Computing Surveys. 53: 1–38. doi:10.1145/3410158
Get In Touch
I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology. Feel free to reach out if you'd like to connect!
Let's Connect
Currently
🚀 Working on exciting new projects
📚 Learning about AI and machine learning
💡 Open to collaborations
☕ Always up for a coffee chat about tech