lauture.bio
reading resume

✓ AfroTech Houston · Nov 2026

We just met at AfroTech Houston.

Thanks for the conversation. Save my contact so this doesn't get lost, or send a quick note so I remember yours.

Save contactSend a note
GitHub LinkedIn

~/ staff platform engineer · atlanta

Marcelin
Lauture

I build the platforms that ML and AI agents run on. Today that is recommendations for Saks and Neiman Marcus.

Career · 10 years

2022

Saks Global

Current

Staff Platform Engineer, Cloud Platform

May 2022 - Present

  • Recommendations up at peak
  • $480K a year saved
  • Safe path for AI agents
  • Agents that do the busywork
  • Company Kubernetes platform

2021

Warner Bros. Discovery

Senior Cloud DevOps Engineer

April 2021 - May 2022

  • Cloud cost forecasting
  • Cost guardrails

2020

Sage Software

Senior DevOps Engineer

January 2020 - April 2021

  • Release pipelines
  • Cloud versions of on-prem products

2016

Aaron's Inc

Infrastructure Engineer, Platform

July 2016 - December 2019

  • Automated server images
  • Pipelines for analytics models

BBA

Kennesaw State University

Bachelor of Business Administration in Information Systems

At Saks Global

Infrastructure that switches off when idle

Product embeddings now run only when the catalog, inventory or content changes.

$480Ksaved every year in ML cloud costs
Requests a second
1,200+
Response time
under ½ s

Recommendations stayed fast with zero downtime through the biggest retail events. Illustrative gauge, not live data.

AI agents that do the busywork

They upgrade dependencies, answer first-line support and triage incidents, and check their own work before an engineer approves. Keeping the predictable steps in plain code cut model usage by 70%.

A safe path for shipping AI agents

Security, reliability and cost controls are built in, so any team can ship an agent the same way it ships a microservice.

Side projects

Waddlecraft

mobile game · coming soon

Your words.
Their hammers.

Pip's colony builds exactly what you tell them to build. The challenge is learning to word it so they build exactly what you wanted.

waddlecraft.com

Spatial Fitting Room

Vision Pro · R&D

Try clothes on your own body, in 3D.

A spatial try-on for Apple Vision Pro. An ML pipeline turns a flat product photo into a 3D garment, fits it to a body-accurate avatar, and bakes a finished piece you can walk around.

  1. Photo
  2. 3D model (TRELLIS)
  3. Fit to body (SMPL-X)
  4. Vision Pro

Resume

Download resume (PDF)

Marcelin Lauture

Atlanta, GA · mlautur1@gmail.com
github.com/mlautur · linkedin.com/in/marcelin-lauture

Staff platform engineer with 10 years building cloud infrastructure at scale, currently leading the ML platform behind recommendations across Saks and Neiman Marcus and the company's framework for production AI agents.

Experience

Saks Global · Staff Platform Engineer, Cloud Platform

Atlanta, GA · May 2022 - Present

  • Owned the stability of personalized recommendations across Saks and Neiman Marcus, closing reliability gaps and readying the ML platform for critical retail events at 1,200+ RPS, <500ms, zero downtime.
  • Cut ML infrastructure costs by $480K a year by moving product embeddings to an event-driven architecture that scales with catalog, inventory and content updates and shuts off when idle.
  • Built the golden path for deploying AI agents, with security, reliability and cost controls built in, so any team can ship an agent the same way it ships a microservice.
  • Reduced platform toil with AI agents that handle code and dependency maintenance, first-line support and incident triage, validating their own changes before an engineer approves, keeping predictable steps in code and cutting model usage by 70%.
  • Leading the design and adoption of the Saks Global Kubernetes platform on EKS, establishing standardized cluster patterns and operational guardrails to enable teams to ship services securely and reliably.
  • Defined developer golden paths and self-service workflows and directed CI/CD modernization across teams, reducing platform toil and accelerating delivery.

Warner Bros. Discovery · Senior Cloud DevOps Engineer

Atlanta, GA · April 2021 - May 2022

  • Shifted cost optimization practices left by implementing prediction tools like Kubecost and Infracost
  • Established cloud cost governance standards by implementing OPA policies and Cloud Custodian

Sage Software · Senior DevOps Engineer

Atlanta, GA · January 2020 - April 2021

  • Designed and implemented end-to-end release cycle pipelines for multiple highly available products
  • Architected and automated cloud environment builds to enable cloud offerings for on-premise products

Aaron's Inc · Infrastructure Engineer, Platform

Atlanta, GA · July 2016 - December 2019

  • Automated machine image creation and management using Packer and CI/CD pipelines
  • Implemented CI/CD pipeline for the enterprise data analytics training models, improving deployment efficiency

Skills

Cloud & Platform:
AWS, Kubernetes (EKS), Terraform, Helm, ArgoCD, Karpenter, Istio, Cilium
ML & AI:
AWS Bedrock, LLM agents, MCP, RAG, LLM evaluation, Kafka, vector search, event-driven autoscaling
Delivery & Observability:
GitHub Actions, CircleCI, GitOps, Datadog, OpenTelemetry, Prometheus, Grafana
Languages:
Python, Go, Bash, TypeScript

Education

Kennesaw State University · Bachelor of Business Administration in Information Systems

Save contactSend a note