I'm Mark Zha, an AI Engineer with a focus on turning unstructured user-generated content into actionable business insights using Python, Java, SQL, and large language models on the Spark platform. I design and refine annotation processes, craft structured prompts for location tagging and classification, and collaborate with product stakeholders to align AI capabilities with business goals. My work emphasizes measurable impact on content discovery and user engagement, with a track record of improving accuracy and reducing manual intervention. I enjoy bridging research and product, leading knowledge sharing, and delivering scalable, efficient solutions. Highlights include achieving 90% classification accuracy on a fine-tuned model, 2 ms latency for sarcasm detection deployments, and building end-to-end data processing pipelines on AWS/SageMaker and Spark/DSPy-based few-shot tuning.

Mark Zha

I'm Mark Zha, an AI Engineer with a focus on turning unstructured user-generated content into actionable business insights using Python, Java, SQL, and large language models on the Spark platform. I design and refine annotation processes, craft structured prompts for location tagging and classification, and collaborate with product stakeholders to align AI capabilities with business goals. My work emphasizes measurable impact on content discovery and user engagement, with a track record of improving accuracy and reducing manual intervention. I enjoy bridging research and product, leading knowledge sharing, and delivering scalable, efficient solutions. Highlights include achieving 90% classification accuracy on a fine-tuned model, 2 ms latency for sarcasm detection deployments, and building end-to-end data processing pipelines on AWS/SageMaker and Spark/DSPy-based few-shot tuning.

Available to hire

I’m Mark Zha, an AI Engineer with a focus on turning unstructured user-generated content into actionable business insights using Python, Java, SQL, and large language models on the Spark platform. I design and refine annotation processes, craft structured prompts for location tagging and classification, and collaborate with product stakeholders to align AI capabilities with business goals. My work emphasizes measurable impact on content discovery and user engagement, with a track record of improving accuracy and reducing manual intervention.

I enjoy bridging research and product, leading knowledge sharing, and delivering scalable, efficient solutions. Highlights include achieving 90% classification accuracy on a fine-tuned model, 2 ms latency for sarcasm detection deployments, and building end-to-end data processing pipelines on AWS/SageMaker and Spark/DSPy-based few-shot tuning.

See more

Experience Level

Expert
Expert
Expert
Expert
Expert
Expert
Expert
Expert
Expert
Expert
Expert
Expert
Intermediate
Intermediate
Intermediate
Intermediate
See more

Language

English
Fluent

Work Experience

Algorithm Engineer at Trip.com Group
March 1, 2024 - Present
Designed, tested, and optimized structured prompts for AI-driven location tagging and classification, reducing manual intervention by 70% and enhancing content discovery. Fine-tuned Qwen-2.5-32b with LoRA to boost classification precision to 90%. Explored DSPy-based few-shot prompt tuning on proprietary models (GPT4-o and Qwen-max) achieving accuracy over 93%. Implemented voting mechanisms across multiple models for 300+ labels with accuracy >95% and external-delivery readiness. Implemented Hive SQL-based data partitioning on HDFS to reduce full-table scans by 65%. Led knowledge-sharing sessions including workshops and conference-style presentations.
Data Scientist Intern at Genex Consulting
August 1, 2023 - November 1, 2023
Performed in-depth analysis on large-scale transportation datasets using Python (Pandas, NumPy), resolving missing values and outliers to improve data quality and forecast accuracy by 20%. Optimized SQL data management, reducing query latency by 15%. Applied ML models to forecast peak congestion periods and optimize routing/scheduling, reducing delays by up to 20%. Designed interactive Tableau dashboards for real-time decision-making.

Education

Master of Information Technology, Major in Cyber Security at University of Melbourne
February 1, 2022 - December 1, 2023
Bachelor of Science, Major in Computer & Software Engineering at University of Melbourne
February 1, 2019 - December 1, 2021

Qualifications

Add your qualifications or awards here.

Industry Experience

Software & Internet, Computers & Electronics, Media & Entertainment