About Me
AI Engineer with 10 years of experience in machine learning, natural language processing, deep learning, data analysis, and delivering production-grade AI solutions. Full project lifecycle experience — from problem defin…
AI Engineer with 10 years of experience in machine learning, natural language processing, deep learning, data analysis, and delivering production-grade AI solutions. Full project lifecycle experience — from problem definition, data preparation, and feature engineering to model training, deployment, visualization, and business rollout. Extensive experience in LLM application development, RAG / Agentic RAG pipelines, prompt engineering, vector databases, knowledge assistants, and enterprise AI solutions. Industry background spans enterprise services, consumer insights, financial services, manufacturing, and HR tech.
Experience
AI Engineer
<p>Translate business requirements into data structures, metrics frameworks, and analytical models, aligning with operational goals.<br>
Develop Python-based data processing pipelines and machine learning algorithms to support business-facing AI solutions.<br>
Build interactive dashboards and reporting tools with Flask to support decision-making for business users and stakeholders.<br>
Design and implement on-premise knowledge assistant systems built on open-source LLMs, Dify, vector retrieval, and RAG.<br>
Contribute to AI solution scoping, proposal writing, and client communication.</p>
Data Scientist
<p>Applied clustering, regression, and classification methods to business modeling, collaborating with engineers on production delivery.<br>
Built data collection, cleaning, preprocessing, and augmentation pipelines to improve model robustness under limited sample sizes.<br>
Fine-tuned pretrained models, coordinated data annotation, and developed sales forecasting models based on review and user data.<br>
Delivered and deployed sales forecasting models for JD.com-related business scenarios.<br>
Built interactive visualizations with Dash to surface patterns, trends, and decision-relevant insights.</p>
Algorithm Engineer
<p>Designed and optimized an NLP-based quality inspection solution for compliance monitoring of financial services sales calls.<br>
Expanded the training set from 5,000+ to 8,000+ samples using SimBERT2 and trained a BERT classifier to detect non-compliant calls.<br>
Compressed the model via knowledge distillation, raising classification accuracy from 84.5% to 92% while keeping detection rate above 90%.</p>
NLP Engineer
<p>Led technical exploration, owning the development, training, optimization, and deployment of NLP and machine learning models.<br>
Built user profiling pipelines combining rule logic, clustering, topic modeling, and predictive tagging to support recommendation applications.<br>
Delivered machine learning solutions across text recognition, image recognition, time series analysis, chatbots, equipment fault detection, and defect detection.<br>
Maintained 100% on-time task delivery, with overall feature completion rate exceeding 90% across owned projects.<br>
Reduced compute and machine costs through model simplification without compromising model performance.</p>
Data Mining Engineer
<p>Applied modeling, forecasting, and algorithmic methods to deliver data-driven solutions across diverse business scenarios.<br>
Authored technical documentation for data processing and label mining based on requirements clarification and feature design discussions.<br>
Built models using R, TF-IDF, Random Forest, and Naive Bayes for research/news topic classification, information filtering, entity extraction, recommendation, asset allocation, and transaction data visualization (Dash dashboards).<br>
Continuously updated the sample library to support model iteration, achieving a final solution accuracy above 90%.</p>