Looking for a job

Kevin Davet

@kevindavet

Message

GenAI Platform Engineer at SAS Institute who took SAS Viya Copilot’s RAG platform from proof of concept to GA.

United States

Message

At SAS Institute, I took the RAG platform behind SAS Viya Copilot from proof of concept to general availability for all SAS customers. I built release-gating evaluation infrastructure, including ground-truth tools for correcting more than 5,000 responses and automated red-teaming alerts.

I’m now architecting a company-wide agent framework at SAS Institute, working with a LangGraph-based agent SDK and Agent2Agent endpoints. I’ve also built domain agents and co-designed the central Go orchestrator for Viya Copilot.

Outside my work at SAS Institute, I was the sole builder of torf, an App Store-approved iOS product for detecting AI-generated media. I developed its models and mobile app, and deployed its cloud-native backend.

Experience

Work history, roles, and key accomplishments

Current

Software Developer, GenAI Platform

Current

SAS Institute

Jan 2026 - Present (9 months)

Architecting the agent framework for customer-facing agentic capabilities across SAS product teams. Built domain agents on the LangGraph-based Viya agent SDK and co-designed the central Golang orchestrator.

LangGraph Agent2Agent (A2A)Go Kubernetes Azure AI Search fastAPI Docker

Sr. Associate Software Developer, GenAI Platform

SAS Institute

May 2025 - Jan 2026 (8 months)

Led technical design and production rollout of the RAG platform behind SAS Viya Copilot from proof of concept to GA. Architected a provider-agnostic RAG framework and built evaluation infrastructure gating releases.

RAG LLM Evaluation Azure AI Search fastAPI Kubernetes Python Lang Chain Red Teaming

Associate Staff Scientist, Computer Vision

SAS Institute

Jun 2023 - May 2025 (1 year 11 months)

Designed Python SDKs providing OpenCV/PyTorch-style APIs over SAS C++ computer vision primitives. Built and deployed automated CI/CD pipelines adopted by over 700 internal repositories.

Python OpenCV PyTorch CI CD GitHub Actions C++

Machine Learning Intern

SAS Institute

May 2022 - Aug 2022 (3 months)

Built a facial-emotion recognition CNN and real-time Event Stream Processing pipelines for live video inference. Extended the CVPy package with image-table join functionality.

Python CNN Event Stream Processing CVXPY

Machine Learning Intern

Cleveland Clinic

May 2021 - Aug 2021 (3 months)

Engineered a CNN and OCR system to classify insurance versus non-insurance cards at 99.3% production accuracy. Conducted demographic and socioeconomic analyses with interactive visualizations.

CNN OCR Python Data Analysis Visualization