Jun 2022 – Jul 2025 · 3 yrs 1 mo · Vancouver, British Columbia, Canada · Remote

• I lead the AI research team and contributed hands-on expertise to enhance legal research. Here's an overview of my key responsibilities:
• 1. Team Leadership and Management:
• Led a dynamic team of over 6 applied scientists and engineers in using NLP and AI to address legal tech challenges, such as sensitive information identification, privileged document recognition, and more.
• Fostered growth within the AI research team, guiding professional development and progression into senior roles.
• Participated in Relativity's Next-Generation Leadership Program to support leadership potential.
• 2. Project Strategy and Collaboration:
• Collaborated with product managers to define project strategies, goals, scope, and deliverables.
• Provided technical mentorship, ensuring project momentum and progress.
• Engaged in cross-functional collaboration to identify customer pain points and drive product innovation.
• Simplified complex technical concepts for non-technical stakeholders.
• 3. AI Research and Achievements:
• Developed over 10 AI models, reducing customer time spent on sensitive information identification by 40%.
• Collaborated with the engineering team to create products like privilege logs and confidential classifiers.
• Introduced AI capabilities like PII identification in spreadsheets using Large Language Models (LLMs), contrastive learning, and more.
• Engineered NER models specialized for legal content, ensuring high precision and recall in entity extraction.

Dec 2019 – Jun 2022 · 2 yrs 6 mos · Vancouver, British Columbia, Canada · Remote

• 1. Designing, implementing, and optimizing advanced NLP models, including BERT, T5, and other transformer-based models, achieving state-of-the-art performance in various challenging tasks, such as:
• Imbalanced document classification
• Legal entity comprehension
• Topic modeling
• Document segmentation
• Named Entity Recognition (NER)
• 2. Contributing significantly to intellectual property with patented NLP-related technologies:
• Provisional patent No. 63/327,989: System and method for intelligent generation of privilege logs.
• Provisional patent No. 63/327,999: System and method for automatic identification of legal entities.
• Provisional patent No. 63/328,005: Systems and methods for self-training a communication document parser.
• 3. Enhancing NLP model interpretability through:
• Statistical techniques, including chi-squared tests and mutual information classification.
• Utilizing gradient-based methods.
• Leveraging tools such as LIME to ensure transparency and trustworthiness.