J
Jesse Quiroga
About
Detail
Houston, Texas, United States
Résumé
Jobs
verified_user
0% verified
Senior AI EngineerNotionDec 2020 - Current (5 years 11 months)• Led development of AI-powered writing assistant using LLMs, improving user engagement by
25% across millions of users
• Designed and implemented Retrieval-Augmented Generation (RAG) system using vector
embeddings and semantic search
• Built scalable prompt orchestration and context management pipelines using Python and
TypeScript
• Developed semantic search platform combining vector search (FAISS) and keyword search
(Elasticsearch), improving relevance by 30%
• Implemented AI workflow automation agents that reduced manual task effort by ~50% for
enterprise users
•
•
Designed evaluation framework to detect hallucinations and improve model reliability, reducing
error rates by- AI EngineerPlaidAug 2019 - Dec 2020 (1 year 5 months)• Built NLP-based transaction categorization system using BERT models, improving
classification accuracy from 78% to 94%
•
Designed real-time fraud detection pipeline using streaming data (Kafka, Spark), reducing fraud
losses by 30%
•
•
•
Developed hybrid ML models (XGBoost + deep learning) for risk prediction and anomaly
detection
Created credit risk scoring API used by financial partners, improving loan decision accuracy by
20%
Reduced reliance on rule-based systems by 60% through scalable ML pipelines
• Deployed high-performance inference services with sub-50ms latency supporting millions of
transactions daily
Education
verified_user
0% verified
- Bachelor of Computer ScienceTexas Tech UniversityJan 2015 - Dec 2019 (5 years)
Projects (professional or personal)
verified_user
0% verified
- AAI Agents for Workflow AutomationMulti-step task automation using LLM reasoning and tool integration
- SSemantic Search EngineHigh-performance hybrid retrieval system for unstructured data
- FFraud Detection SystemReal-time ML system processing millions of financial events daily
- LLLM-Based Writing AssistantBuilt production-grade generative AI feature with prompt engineering and retrieval pipelines