Contractual devlopment work on hard problems in ai data and infrastructure.

• Designed and implemented overall solution architecture in AWS cloud• Built initial NLP product via custom built domain specific decision trees with semantic interconnections• Later migrated to AI based NLP utilizing LLMs for NLU and NLG as needs dictated• Analyzed LLM performance in different contexts for pharmaceutical education• Utilized GPT 2-4.5 and Llama2 as performance and security needs dictated• Architected and implemented Python microservices using copilot to support thousands of requests per minute• Built CICD pipeline to quickly deploy changes pushed to Github directly to our AWS services• Developed MySQL database structure, data architecture, and REST interfaces for patient data• Created comprehensive testing methodology to minimize patient-facing bugs and build customer confidence• Evaluated VecDB options for RAG including pinecone and vectorized SQL• Deployed comprehensive LLMops solution with Vellum to speed up development speed and rapidly deploy• Created and assigned two-week sprint tasks to continuously drive agile projects forward• Mentored intern on software development and data analysis by overseeing progress and guiding work direction• Identified and implemented organization-wide compliance policies and architectures for Soc2 and HIPAA• Researched, presented and resolved legal and regulatory issues related to product development and FDA approval

• Developed system dynamic economic model using Stella an all in one system dynamic modeling software• Utilized model for prediction of Latin American bond yields based on historical yield rates, and outstanding debt• Analyzed and interpreted historical data to calibrate technical model• Built python dashboard for data visualization that combined historical yield rates from Bloomberg with model predicted future yield rates

• Reworked communication and diagnostic protocols of printer system leading to 50% faster debugging• Separated development and production environments to speed development and prevent bugs• Developed and implemented automatic self-calibration process to reduce customer site visits• Improved deployment and versioning process to reduce problem identification times• Identified hardware/software interface issues to assist in debugging customer systems• Added configuration GUI allowing customers to make use of previously inaccessible printer options• Maintained firmware for 4 different specialized microprocessors

Originally Blink for home, a small company that was later acquired by amazonDesigned and codified procedures for product manufacturing and deliveryWorked with Operations to design and improve RMA proceduresDesigned and updated multiple testing setupsDiagnosed and troubleshot various customer issuesControlled regression testing of firmwareInterfaced with manufacturing factories for firmware deliveryDeveloped audio firmware for custom processorImplemented AGC algorithm that minimized resource usageManaged teams of testers to verify functionalityDesigned several new product tests for manufacturing verificationDeveloped product verification tests