About Me
I’m Wen Qing, a machine learning engineer based in London with 7+ years of experience building ML systems across HR Tech, FinTech, AdTech, and food delivery. I like building solutions that are not merely technically sophisticated, but also reliable, transparent, and meaningful to the people who use them.
What I do
- LLMs & NLP: fine-tuning LLMs and embedding models, entity matching over messy text, and LLM-as-judge evaluation pipelines. At SilentEight, fine-tuning on synthetic data improved recall by 44%.
- Production ML: end-to-end ML in Python, from research to deployment, with PySpark, AWS (Bedrock, SageMaker), GCP, Docker, Airflow, and MLFlow. My work has saved thousands of manual hours across Health Tech, Quick Commerce, Social Media, and Reg Tech.
- Explainable and observable AI: failure taxonomies, human-in-the-loop feedback, and monitoring that tells a team what to fix next.
- Leading and communicating: mentoring other engineers, leading research teams, and telling the story of technical work to stakeholders. I thrive in innovation-driven, start-up style environments.
What I’m into
- Music and AI. I have an MSc (Distinction) in Sound and Music Computing, and my thesis on hierarchical music generation was published at CMMR. Projects along the way include multi-singer separation and a multimodal creative system. I also run The Orchestral Pit, a daily feed of new music information retrieval research.
- Music industry problems. I lead data science at Octavate, using ML to tackle A&R, such as predicting breakout artists.
- Building things. Side projects, personal tech, and weekend experiments.
- Writing. I write about music and ML and career lessons on Coding in Tune, and about travel adventures in the real world.
Let’s chat
Always up for a conversation about evaluating LLM systems, applied ML, the music industry, or the occasional philosophical debate about whether computers can truly achieve creative autonomy. If you’re interested in collaborating, let’s connect!