Prior professional work · 2020–2023

Technical risk reduction across a software investment portfolio.

As a Senior Platform Engineer operating as an internal technical consultant, Brandon entered portfolio companies where an immediate infrastructure problem was inseparable from a larger business constraint.

The mandate was to understand the system, make the risk legible, and turn technical investment into an operational result—not simply administer cloud resources.

Organization: AZ Crown Investments
Role: Senior Platform Engineer · Internal Technical Consultant
Principal engagements: Trovo.ai · QuickSpark

The portfolio mandate

Different systems. The same underlying question.

What technical risk was preventing the business from moving forward?

AZ Crown deployed Brandon into companies carrying significant technical risk or strategic opportunity. The work crossed cloud and platform architecture, data systems, infrastructure automation, access governance, delivery practice, observability, cost, and communication with business leadership.

At Trovo, the immediate threat was economic. At QuickSpark, it was accumulated delivery, operational, and organizational debt. In both cases, progress began with evidence rather than a preferred technology.

Engagement 01 · Trovo.ai · approximately 2020–2021

When platform economics become business risk.

Trovo operated a multi-partner data cooperative. Its AWS Glue-heavy processing model worked technically, but its economics had become unsustainable as the company continued onboarding partners.

Historical cloud expenditure was approaching approximately US$1 million annually. Sorting, aggregation, and grouping workloads were consuming enough Glue capacity that infrastructure cost threatened the viability of the business.

Brandon investigated alternatives and formed the view that Snowflake could perform the analytical and transformation work more economically. Instead of asking the company to accept a wholesale migration on theory, he selected a representative production workload and built a proof of concept.

The measured result established technical feasibility and the economic case. That production evidence secured support for a broader migration, which Brandon then led.

Established outcomes

A more sustainable data platform.

approximately US$900,000: in annual infrastructure savings
100+: Snowflake databases administered across approximately 40 partners
approximately 15 PB: in the broader governed multi-partner data estate
Hours: for partner onboarding that had previously taken weeks or months

Platform & governance

Migration was only part of the work.

The resulting platform combined Amazon S3, Amazon SNS, Snowpipe, Snowflake, Kubernetes/EKS, Terraform, and custom AWS automation. S3 events flowed through SNS into ingestion or processing pipelines. Snowpipe supported continuous ingestion, while one-off EKS workloads handled specialized processing without keeping every form of compute permanently active.

Brandon also administered the Snowflake environment: creating and sizing virtual warehouses against business service-level objectives, tuning workloads before adding compute, applying auto-suspend where appropriate, and monitoring spend and long-running queries.

Least-privilege roles and grants, access conventions, and database and object naming standards established clearer governance. The naming and organization model preserved the relationship between information and its contributing partner instead of allowing provenance to disappear into an undifferentiated data lake.

Infrastructure was managed with Terraform and checked routinely for drift. The approximately 12-account AWS estate used AWS Organizations, identity-provider-based SSO, MFA, least-privilege controls, and Teleport-based access.

Engagement 02 · QuickSpark · approximately 2021–2022

Modernize delivery before attempting a rewrite.

QuickSpark’s mission-critical lending application and its surrounding delivery model reflected practices dating largely from the mid-2000s. Small changes could take months, regressions were common, and important knowledge and authority were concentrated in one position.

Brandon began with the engineering system around the application. He trained employees in Agile-aligned practices, introduced stronger testing and version-control discipline, and connected source control to project and change-management workflows.

Changes gained technical context, business justification, and auditable history. Technical and nontechnical stakeholders could work through a shared process, reducing key-person dependence and making changes easier to approve, trace, and troubleshoot.

Infrastructure & visibility

Move the platform—and make its behaviour visible.

Brandon migrated the primary in-house lending application from on-premises infrastructure to AWS in 87 days. The move reduced dependence on physical infrastructure and gave the company greater delivery flexibility.

He implemented centralized observability using Elasticsearch, Logstash, Kibana, and Graylog. Access logs were ingested and normalized through autoscaling Graylog instances, creating an audit trail that made production behaviour traceable.

That visibility enabled the team to diagnose longstanding production defects—including issues that had persisted for years—without reconstructing events from incomplete evidence.

Brandon also designed a progressive microservice-modernization strategy. The legacy system could continue serving the business while capabilities were separated or replaced incrementally, avoiding an all-at-once rewrite and lengthy feature freeze.

Outcome

A safer path for ordinary change.

The work reduced bugs, outages, and dependence on individual memory while improving collaboration, traceability, delivery flexibility, and the company’s security posture.

Just as importantly, modernization no longer depended on pausing the business for a speculative rewrite. The organization had a practical path for improving the system in stages.

The common method

Begin with the business risk.

The tools differed. The operating pattern remained consistent across both engagements.

Investigate the system’s actual technical and economic behaviour.
Establish visibility and traceability before making broad claims.
Test major architectural decisions through practical production evidence.
Connect technical findings to cost, delivery, security, and business continuity.
Introduce repeatable controls, automation, and operating conventions.
Modernize incrementally when a wholesale rewrite would increase risk.

Provenance

Prior professional work, with attribution intact.

Brandon was employed by AZ Crown Investments, not directly by Trovo.ai or QuickSpark. Historical scale and spending figures are approximate recollections and are labelled as such. This work predates Emberside Consulting.

This case study intentionally excludes unverified architecture, compliance, uptime, transaction-volume, and performance claims.

Discuss a similar constraint