Reducing inference latency by 40 percent across personalized feed algorithms at LinkedIn set the stage for my work in real time ranking systems. By refactoring our PyTorch deployment pipeline, I handled a daily volume of 50 million requests while ensuring high availability for critical user engagement features. This technical challenge taught me how to balance complex model architectures with the rigorous uptime requirements of a global social platform. My background in building scalable machine learning services positions me to contribute immediately to the sophisticated matching technology powering the ecosystem at ZipRecruiter.
At Yelp, I leveraged XGBoost to optimize local search relevancy, which improved click through rates by 12 percent within three months of deployment. During my tenure at LinkedIn, I developed a custom Scikit Learn preprocessing pipeline that streamlined data ingestion, cutting model training time from ten hours down to two. I also utilized Hugging Face Transformers to implement advanced sentiment analysis for community moderation tools, boosting detection of spam by 25 percent. These projects relied on Python and Apache Spark to maintain data integrity across massive, distributed data clusters used in daily production environments.
ZipRecruiter occupies a unique position in the employment sector by prioritizing algorithmic efficiency to bridge the gap between job seekers and hiring managers. My specific experience in ranking optimization and large scale feature engineering aligns perfectly with the goal of improving your matching engine performance. My commitment to transparent, interpretable machine learning models mirrors your company approach to fair hiring practices. By integrating advanced deep learning techniques with the robust data infrastructure you maintain, I aim to push the boundaries of how efficiently your users navigate the job search experience.
My combination of industry experience and technical expertise in the stack used by your engineering team ensures I can deliver impact within your organization. I have navigated the complexities of production level machine learning and look forward to discussing how my work at LinkedIn and Yelp translates into results for your technical initiatives. Please reach out to schedule a time for us to evaluate how my skills will help advance your current projects. I appreciate your consideration of my profile for this position and look forward to hearing from you.
Best regards,
Patrick Sullivan