1,300×+
Production throughput improvement
A critical flow moved from approximately 0.018 to 24 operations per second after diagnosis and redesign of the critical path.
Principal ArchitectIrish-born · Hangzhou-based
Jay Croghan
Principal Architect for complex systems that cannot afford to fail.
I modernise enterprise financial platforms, solve the performance problems others avoid, and build practical AI systems that make engineering teams faster without giving up control.
Current
Principal Architect
SS&C Technologies
01 / Selected impact
Measured outcomes, not architecture theatre.
The numbers matter because each one represents a difficult system understood well enough to change safely.
57,000
Policies handled by a redesigned flow
Repetitive application behaviour became a set-based SQL and parallel C# processing pipeline designed for the actual workload.
4,600+
Changesets analysed through governed AI
TFVC history, review bundles, model analysis and actionable reports combined in a controlled engineering workflow.
50+
Technical professionals led or mentored
Architecture, development and QA across globally distributed teams—with technical direction kept close to delivery.
02 / About
Architecture that survives contact with production.
Close to code. Close to data. Accountable for outcomes.
I am a hands-on architect, not a slide-deck-only architect. I work where architecture, performance, delivery and production reality meet.
Most of my career has been spent on complex, regulated financial platforms across pensions and investments. I am most useful when a system is large, established, heavily integrated, difficult to change and under real delivery pressure.
I can move from an executive or client conversation into platform strategy, low-level design, SQL execution plans and working implementation without losing the thread. That range matters: difficult systems rarely fail inside one neat layer, and the right answer has to work for the business, the engineers and production.
03 / Case studies
Three systems. Three kinds of leverage.
Sanitised examples of diagnosis, redesign and governed automation. No client names or confidential architecture.
Case 01
0.018 → 24
operations / second
Production performance rescue
Find the constraint. Redesign the path.
A production process had slowed to approximately one completed operation every 55 seconds. Treating it as a system problem—not a single slow line of code—I traced behaviour across the critical path, separated symptoms from constraints and identified where work was being amplified.
The solution changed how the flow executed rather than applying a local patch. It removed the dominant bottleneck, protected the surrounding platform and restored throughput to approximately 24 operations per second.
Result More than a 1,300-fold throughput improvement.
Case 02
57,000
policy workload
High-volume processing redesign
Put each kind of work in the right layer.
The original flow repeated too much work in the application layer and paid the cost on every policy. I redesigned it around set-based SQL, table-valued parameters and parallel C#, moving each kind of work to the part of the stack best equipped to perform it.
The outcome was not simply “faster code.” It was a clearer data and execution pipeline with fewer round trips, controlled concurrency and behaviour that could be reasoned about under load.
Result A deliberate, high-volume data and execution pipeline.
Case 03
4,600+
TFVC changesets
AI-assisted engineering
Use models inside a controlled engineering system.
I built a structured pipeline around TFVC history and code review. It collected changes, assembled review bundles with the context a model needed, invoked model-based analysis and produced reports that engineers could act on.
The workflow treated AI as an auditable component: bounded inputs, inspectable outputs and human technical judgment at the decision point. It is the kind of AI adoption I trust—useful, controlled and attached to a real delivery problem.
Result Structured review evidence rather than ungoverned model output.
04 / Experience
Twenty years of software, from implementation to platform direction.
The timeline follows the exact current LinkedIn dates for the principal architecture roles, resolving the old site’s SS&C / Insigma overlap.
-
SS&C Technologies
Principal Architect
- Own solution and platform architecture for a regulated investments and pensions platform.
- Connect client requirements, modernisation strategy, detailed design, engineering delivery and production operation.
- Lead and mentor more than 50 architects, developers and QA professionals across globally distributed teams.
- Stay active in performance, production, integration and AI-assisted engineering work that crosses technical boundaries.
-
SS&C Technologies
Senior Technical Architect
Led solution, integration and low-level design for enterprise financial-platform change, translating difficult requirements into implementable work across architecture, engineering and QA.
-
Independent work
Career break — personal goal pursuit
Independent technical work, continued software experimentation and creative opportunities in China.
-
Insigma Hengtian Software Ltd.
Senior Architect
Directed and supported technical leads designing a modern enterprise fintech platform, connecting high-level architecture, detailed design and implementation across an international delivery environment.
-
DST Systems
Technical Architect
Guided technical leads and engineering teams on a modern enterprise financial-services platform, working across solution structure, integration boundaries, detailed design and delivery support.
-
Loki Hostel / Loki Hostels
Software Developer
Built public web products and operational software, including the main Loki Hostels site and LokiPlanet, using Ruby on Rails and related web technologies.
-
Percana
Senior Software Developer
Built and supported an enterprise financial-services suite, developing deep experience in C#, .NET, SQL Server, desktop systems, service integration, production support and performance-conscious implementation.
05 / Selected independent work
The same curiosity, outside the enterprise boundary.
A restrained selection that demonstrates low-level engineering, product range and practical experimentation.
01
Game-engine modification
ZombieMod:Source
Created a widely adopted Counter-Strike: Source modification using low-level engine hooks and patches. The work required reverse engineering, native C++ and the discipline to keep a community-run system stable as the underlying game evolved.
C++ / engine hooks / reverse engineering
02
Medical-device software
MIRA
Built a standalone C# control application for a medical laser device, integrating medical imaging, intelligent configuration and Bluetooth LE communication between the software and device.
C# / imaging / Bluetooth LE
03
Enterprise knowledge and AI
Retrieval, review and model evaluation
Built retrieval over an approximately 11,000-page knowledge base, evaluated self-hosted and frontier models, automated code-review analysis and operated practical self-hosted AI infrastructure.
RAG / model evaluation / governed automation
06 / Technical range
Broad enough to see the system. Deep enough to change it.
Architecture and modernisation
Solution and platform architecture, legacy modernisation, high- and low-level design, standards and delivery strategy.
C# and .NET
C#, .NET Framework, Windows desktop software, Windows services, WCF, concurrency, integration code and hands-on implementation.
SQL Server and performance engineering
Set-based design, table-valued parameters, execution plans, data pipelines, bottleneck diagnosis and production throughput.
Distributed and integrated systems
Service boundaries, systems integration, TCP/IP, SOAP, platform dependencies, fault diagnosis and supportable delivery.
AI-assisted engineering and model evaluation
RAG, code-review automation, model orchestration, open and frontier model evaluation, context design and human-in-the-loop governance.
Linux, infrastructure and self-hosting
Linux administration, Apache/Nginx, networking, Bash, self-hosted applications, security hardening and practical operations.
Technical leadership and client engagement
Client conversations, architecture mentoring, globally distributed delivery, technical direction and translating constraints into decisions.
07 / Contact
Have a difficult system? Let’s talk.
I am interested in architecture leadership, platform modernisation, performance-critical systems and technically ambitious product work—especially where senior judgment still needs to stay close to engineering reality.