Scalexa Consulting / selected work

Published technology work, with the evidence in view.

Review six source-linked records: Be-IQ infrastructure rescue, Cityblock Health data unification, Pangea Bio's SimMS GPU engine, Mozilla Foundation fundraising infrastructure, ListenLayer AWS DevOps and security, and Makro's retail MarTech platform.

Book a Free 30-Min Call

Six engagements, clearly scoped

Different technical risks. The same standard of evidence.

These records use published client, challenge, approach, outcome, and proof details only. Metrics are labelled source-reported, and testimonials appear only where approved wording is published.

Be-IQ / AWS infrastructure rescue

DevOps & Cloud Migration

Published case study / client named

Read published source
Be-IQ AWS infrastructure migration case study

Context

Be-IQ engaged Scalexa after failing EKS infrastructure left its AWS environment in need of a production-grade rebuild around ECS, CI/CD, and MongoDB.

The challenge

Deployments were slow and error-prone, MongoDB lacked proper backup and scaling strategies, and the platform needed reliable operations without sacrificing rollback safety.

Rebuild the path to dependable delivery.

  1. 01

    Move from EKS to ECS

    Rebuilt the AWS environment around ECS to replace the failing infrastructure path.

  2. 02

    Build robust CI/CD

    Established automated, zero-downtime deployments with rollback capability.

  3. 03

    Add MongoDB resilience

    Implemented MongoDB hosting with automated backups and monitoring for stronger recovery readiness.

Source-reported outcomes

99.9%+ uptime, deployments reduced from hours to under 10 minutes, and zero-downtime rollback capability.

“Hiring Scalexa was one of the best decisions we made! They have built us a fully scalable, robust architecture and steered us through every step of the complex process. Not only have they delivered a vastly improved and stable hosting architecture, they also massively upgraded our security and with it my peace of mind! Scalexa has encyclopedic knowledge and understanding of Technology; they are strong communicators and, most importantly, deliver!”
Michael FreeFounder, BeIQTestimonial source

Cityblock Health / data unification

Data Engineering

Published case study / client named

Read published source
Cityblock Health data engineering case study

Context

Cityblock Health needed its Amazon Redshift database and Google Cloud storage unified in an ETL pipeline that could support Apache Beam and future ML/AI work.

The challenge

The migration had to preserve 100% data integrity while working within strict budget and timeline constraints across previously incompatible systems.

Unify the data path without losing precision.

  1. 01

    Centralise in BigQuery

    Implemented BigQuery as the central data warehouse and unified repository.

  2. 02

    Orchestrate with Apache Beam

    Built Apache Beam pipelines for scalable data processing and ETL orchestration.

  3. 03

    Prepare the ML/AI path

    Delivered a futureproof ETL foundation with ML/AI capabilities across the unified data.

Source-reported outcomes

Zero data loss during migration—every decimal preserved—and delivery under budget and ahead of schedule.

Pangea Bio / SimMS GPU Engine

High-Performance Computing & Bioinformatics

Published case study / client named

Read published sourcePeer-reviewed Bioinformatics publication
Pangea Bio SimMS GPU engine case study

Context

Pangea Bio needed a GPU-accelerated cosine-similarity engine for mass-spectrometry workloads and repository-scale searches.

The challenge

The engine had to preserve algorithmic agreement with MatchMS while turning searches that took tens of CPU-days into a practical workflow.

Accelerate the reference implementation faithfully.

  1. 01

    Design parallel GPU kernels

    Designed parallel GPU kernels for greedy and modified cosine similarity.

  2. 02

    Preserve algorithmic equivalence

    Matched MatchMS across all tolerance ranges with 99.99% agreement against the reference implementation.

  3. 03

    Validate at repository scale

    Reduced repository-scale searches from tens of CPU-days to hours on a single GPU.

Published outcome

~1000× acceleration over MatchMS, 99.99% agreement with the reference implementation, and peer-reviewed publication in Bioinformatics, Volume 41, Issue 3, March 2025.

Mozilla Foundation / fundraising infrastructure

Nonprofit Technology

Published case study / client named

Read published source
Mozilla Foundation fundraising infrastructure case study

Context

Mozilla Foundation needed robust, scalable fundraising infrastructure for high-volume donation campaigns, multiple currencies, payment methods, and donor management.

The challenge

The platform had to remain reliable during critical end-of-year campaigns while giving the team useful campaign performance visibility.

Make fundraising infrastructure dependable at campaign scale.

  1. 01

    Build high-availability processing

    Developed fundraising infrastructure with auto-scaling for donations at scale.

  2. 02

    Support sustained giving

    Added recurring donation management and a donor portal for subscriptions and history.

  3. 03

    Harden campaign operations

    Used blue-green deployments and load testing for critical fundraising periods.

Source-reported outcomes

Donations processed from 100+ countries, 99.99% uptime during critical campaigns, and a reported 40% reduction in payment-processing failures.

“Our experience with Gareth was great - he onboarded onto a project quickly and was able to help us. Looking forward to working with him more in the future.”
Sarah SwayzeEngineering Director, Mozilla FoundationTestimonial source

ListenLayer / AWS DevOps & security

DevOps & Cybersecurity

Published case study / client named

Read published source
ListenLayer AWS DevOps and security case study

Context

ListenLayer needed reliable, secure AWS infrastructure for a marketing data platform processing millions of events daily while meeting enterprise privacy and compliance requirements.

The challenge

The platform needed elastic scale, real-time performance, zero-trust security, and a deployment process that could support faster releases without compromising operational control.

Build secure infrastructure for real-time data operations.

  1. 01

    Design highly available AWS infrastructure

    Deployed ECS with Fargate, RDS Multi-AZ, ElastiCache, CloudFront, Lambda, and Terraform-based infrastructure as code.

  2. 02

    Modernise delivery practices

    Implemented GitHub Actions CI/CD, blue-green deployments, automated security scanning, and centralised CloudWatch observability.

  3. 03

    Apply defence-in-depth controls

    Added AWS WAF, private subnets, least-privilege IAM, Secrets Manager, and encryption at rest and in transit.

  4. 04

    Operationalise compliance monitoring

    Implemented SOC 2 Type II controls, GDPR and CCPA handling procedures, AWS Config rules, CloudTrail, and real-time alerting.

Source-reported outcomes

99.99% uptime, zero security incidents since deployment, sub-100ms event-processing latency, SOC 2 Type II certification achieved and maintained, releases increased from weekly to multiple daily, and infrastructure costs optimised by 40%.

Makro / retail MarTech & data platform

Retail MarTech & Data Platform

Published case study / client named

Read published source
Makro retail marketing technology platform case study

Context

Makro needed a digital marketing platform to unify web, email, and social channels while providing customer profiling and behavioural insight across its large retail catalogue.

The challenge

The platform had to support data-driven marketing, real-time customer insight, targeted promotions, and intent capture across a catalogue of 55,000+ products.

Unify retail marketing data around customer intent.

  1. 01

    Connect the marketing channels

    Unified web, email, and social marketing channels through one interface for managing digital campaigns.

  2. 02

    Build customer insight capabilities

    Implemented customer profiling, dynamic segmentation, and real-time behavioural tracking and event capture.

  3. 03

    Capture product intent

    Added wishlist functionality to capture customer intent across 18,000 product listings and support more relevant communication.

Source-reported outcomes

4 million monthly page views from 500,000 unique visitors, unified management of web, email, and social channels, and customer intent data captured across 18,000 products within a 55,000+ product catalogue context.

Start with the evidence

Bring us the technical problem that needs a clearer next move.

Tell us what is changing, what is stuck, and what a useful next move needs to make possible.

Book a Free 30-Min Call