Open to opportunities

Jason Chou

@jasonchou

Message

Senior data scientist delivering NLP, semantic search, and production ML solutions.

United States

Message

What I'm looking for

I’m looking for a team where I can ship production NLP and semantic search systems end-to-end—modeling, evaluation with XAI, and scalable pipelines—partnering with stakeholders in Agile while owning measurable performance and latency improvements.

I’m a data science and machine learning professional with 7+ years delivering NLP, semantic search, and entity resolution for document-centric analytics. I thrive on turning mission requirements into prioritized AI and analytics use cases with real stakeholder impact.

In my work, I’ve designed transformer-based retrieval and ranking models in PyTorch and HuggingFace, improving precision@10 by 22%. I’ve built OCR/ICR-driven ingestion pipelines that processed 5M documents and reduced parsing errors by 47%, and I’ve engineered end-to-end semantic search integrated with Elasticsearch and Databricks to cut median query latency by 40% and boost recall by 30%.

I also focus on production readiness and measurable trust: I’ve implemented graph-based entity resolution with Neo4j to reduce duplicates by 78% across a 20M-entity index, deployed models via REST APIs and scheduled batch pipelines using Docker, Airflow, and cloud services, and used SHAP/LIME-based explainability to increase XAI adoption by 65% across pilots.

Experience

Work history, roles, and key accomplishments

Senior Data Scientist / ML Eng

Integer Group

Sep 2021 - Mar 2026 (4 years 6 months)

Delivered NLP and semantic search pilots with 10+ stakeholders, improving precision@10 by 22% using transformer-based retrieval and reducing median query latency by 40% while boosting recall by 30%. Built end-to-end OCR/ICR ingestion for 5M documents (47% fewer parsing errors), enabled entity resolution to cut duplicates by 78%, and shipped 12 Tableau/Looker dashboards that reduced analyst time-to

PyTorch Faiss Elasticsearch OCR ICR Pipelines Spark Delta Lake Airflow SHAP LIME Explainability

Data Scientist / ML Engineer

Integer Group

Feb 2019 - Aug 2021 (2 years 6 months)

Partnered with product owners and data engineers to scope and run reproducible Databricks experiments, contributing to 6 pilot initiatives. Improved NER recall by 18% and precision by 12%, reduced data freshness from 48 hours to 4 hours with Airflow/Spark scheduling, and built a 2M-page labeled OCR corpus for downstream NLP models.

Databricks Pyspark Pandas scikit learn TensorFlow Apache Airflow REST Container Packaging

Data Analyst Intern

Divercety LLC

May 2017 - Aug 2017 (3 months)

Cleaned and analyzed marketing datasets using R and SQL, reducing missing-value rates by 27% across key tables. Built NLP preprocessing with tokenization and TF-IDF for a 30,000-document corpus and produced 4 Tableau/Matplotlib reports for marketing and analytics teams.