Fractional Technical Operator

I take over the systems your business cannot afford to have fail.

AI Systems · Cloud Architecture · Product Engineering · Lead Systems

I step into production data platforms, cloud environments, and AI initiatives that are too important to leave unowned — and take direct responsibility for getting them right, from architecture through deployment.

Discuss a Project View the Work

Tampa Bay, Florida · Available for select project engagements

$Brian Lapuste, fractional technical operator, in a dark suit in front of a city skyline$

500K+
healthcare claims processed daily
169M+
U.S. addresses in platform scope
30K+
workforce users on federal SSO
150+
unnecessary AWS resources removed
6
GovCloud apps security-hardened

AWS

Claude

OpenAI

Next.js

React

SQL

Java

Python

Spark

Most technical failures aren't about missing tools.

They're about ownership. A data pipeline the whole business depends on that nobody fully understands anymore. A cloud environment that grows more expensive and less legible every quarter. An AI initiative that impressed in the demo and stalled on the way to production.

These problems persist not because the technology is hard to buy, but because no one is accountable for the system end to end.

I take ownership of the problem — architecture, implementation, and the system in production — not just a deliverable.

Proof

Systems I've owned under real stakes.

01 — Regulated Healthcare Data Platform

Navitus Health Solutions

Senior-Level Data Engineer

500,000+ claims processed daily.

The problem

A daily healthcare-claims ETL pipeline was fragile and error-prone, supporting regulated medical data workflows the business depended on.

What I owned

Owned the pipeline rebuild at scale using AWS Glue, Step Functions, and Terraform, with data-quality checks built in — and owned deployment execution, release coordination, and resolution of deployment errors across Azure DevOps.

Outcome

Removed 150+ unnecessary AWS resources, improving maintainability, cost awareness, and long-term platform stability.

AWS Glue
PySpark
Python
Step Functions
Terraform
AWS CDK
S3
Athena
ECS
EMR
Azure DevOps

02 — National-Scale Logistics & Cloud Cost Ownership

United States Postal Servicevia ECS Federal

Senior Full-Stack Engineer – DevSecOps (Government Security Clearance)

Platform operating across 169M+ U.S. addresses, Continental US and Hawaii.

The problem

Nationwide carrier delivery data needed reliable large-scale ingestion, and cloud compute needed cost-efficient purchasing without sacrificing reliability — on a platform operating within the scope of 169M+ U.S. addresses.

What I owned

Designed 6TB cron-based Hadoop batch ingestion pipelines on EMR processing millions of daily breadcrumb delivery records; refactored 150+ REST endpoints and rebuilt the Java DAO architecture with service-layer segregation and structured exception handling; optimized ECS/EMR purchasing strategy from On-Demand to Spot and Reserved capacity.

Outcome

Reduced unhandled exception rates significantly through reusable interfaces and structured error handling across distributed systems.

Separately appointed Lead DevSecOps Engineer for CVE remediation and compliance across six GovCloud applications — security hardening, IAM lifecycle controls, disaster recovery via read replicas, and infrastructure modernization with AWS CDK. Specifics stay confidential, as they should for sensitive federal systems.

AWS
Java
Spring Boot
Angular
Python
EMR
Hadoop
ECS
S3
DynamoDB
RDS
AWS CDK
CloudFormation
PowerShell

03 — Enterprise Referral Data Platform

Cigna Evernorth

Software Developer – Business Analytics

500GB+ daily enterprise data workloads; 200,000+ medical referrals processed through the pipeline.

The problem

Medical referral processing ran through fragmented legacy and modern systems — mainframe JCL, Teradata, PostgreSQL — with throughput constraints.

What I owned

Engineered Python-based serverless and stateful AWS systems (Lambda, VPC, S3, Glue, API Gateway); built RESTful APIs with Spring Boot integrating the Java backend and React frontend; implemented infrastructure as code with Terraform.

Outcome

Modernized the referral pipeline across legacy and cloud systems, easing the throughput constraints on 200,000+ referrals.

AWS Lambda
S3
Glue
API Gateway
VPC
Terraform
CloudFormation
Java
Spring Boot
React
PostgreSQL
Teradata
Talend ETL

Approach

How I work.

Diagnose

Understand the system as it actually runs — where it's fragile, what it costs, and who depends on it.

Architect

Design the target state with the real constraints named: compliance, cost, team, and timeline.

Build & Own

Implement it directly — infrastructure as code, tested pipelines, honest error handling — and own the deployment.

Operate

Stay accountable in production until the system is stable, documented, and genuinely handed over.

Credentials

Verified, not claimed.

AWS Certified DevOps Engineer – ProfessionalDOP-C02

Amazon Web Services Training and Certification · Earned February 2025 · Valid through 2028

Verify on Credly

AWS Certified Solutions Architect – AssociateSAA-C03

Amazon Web Services · Earned May 2023 · Listed validity through 2026

Verify on Credly

B.S. in Computer ScienceCum laude · GPA 3.6 · Worked full-time throughout

University of South Florida · Graduated December 2022

About

Brian Lapuste

Brian Lapuste is a Tampa Bay–based technical operator who has owned production systems in regulated healthcare and federal logistics environments — from data pipelines processing hundreds of thousands of records daily to national-scale cloud infrastructure. He works project by project, taking direct ownership of the system rather than handing off a deliverable.

Contact

If the system is important enough to worry about, it's worth a conversation.

Tell me what you're trying to solve. If it's a fit, we'll scope it as a project with clear ownership — and if it isn't, I'll say so.

Connect on LinkedIn Download Résumé