Rex Asabor

Led the launch of mobile app’s rebrand, syncing engineering, design and marketing functions.

Built a framework to enable A/B testing and real-time evals of our fine-tuned LLMs.

Developed data pipelines and key infrastructure for a LLM-based recommendation engine.

Used prompt engineering to maximize quality & minimize cost of inference at >100k WAU scale.

Researched cost and inference optimization techniques for deploying large language models.

Collaborated on data collection strategy for training a proprietary LLM.

Reported to the CTO as engineer #9, as we grew headcount 3x and revenue 200x in 3 months.

Spearheaded early trust & safety engineering efforts; deployed content moderation algorithms.

Enhanced our account verification process by leveraging ML models and internal tooling.

Prototyped an item-to-item recommendation engine with a small team of engineers.

Offered support for rendering 11 new file types in our 3D environments.

Developed integrations with Google Drive, OneDrive, and more for file imports.

Proposed & led a partnership with Miro to transfer sticky notes between apps.

Leveraged early GPT-3 beta access to populate 3D meeting rooms with relevant content.

Built a new revenue dashboard and API for Snapchat Partners to analyze their ad revenue.

Prototyped a CV algorithm for automatically inserting ads into Snapchat Discover stories.

Developed new ML-based algorithms & tools for dataset deduplication and entity recognition.

Achieved a 13% increase in labels for key multi-terabyte alternative text datasets.

Developed a new algorithm to identify mouse strain genes, increasing recall by 6%.

Invited to present at Harvard Medical School and the Asian Pacific Bioinformatics Conference.

Senior Capstone: Built infrastructure auto-scaling algorithms via k8s & Kafka for Crowdstrike

President’s Scholar ($250k), Lockheed Martin STEM Scholar ($40k), Neo Scholar, Conviction Fellow