Senior backend & platform engineerToronto · open to the right senior engineering role
Building reliable backend platforms, applied AI systems, and independent products.
I'm Marshall, based in Toronto. I bring more than a decade of Java and distributed-systems experience across merchant, catalog, and order domains, designing and delivering high-concurrency, low-latency, highly available systems that support 1M+ orders per day. I have also built event-driven Kafka pipelines processing 300K+ events per second at peak and 1B+ events per day, with hands-on experience in Spark and large-scale data processing. Recently, I have focused on AI automation workflows, including AI harnesses, RAG, and agent collaboration.
- Backend & distributed systems
- Platform & production delivery
- Retrieval, RAG & evaluation
- Java · AWS · TypeScript · Python
How I engineer
Rigor should make delivery clearer, not heavier.
The tools change by domain. These four constraints stay useful across backend platforms, AI systems, and independent products.
Correctness before cleverness
Make invariants, ownership, idempotency, and failure behavior explicit before optimizing the happy path.
Observable by design
A change is not complete until its state, evidence, and recovery path can be inspected.
Evaluation over demos
For AI systems, separate retrieval, grounding, task success, latency, and refusal instead of trusting one fluent answer.
Reversible delivery
Prefer staged rollout, readback, and bounded side effects so speed does not depend on optimism.
Career arc
A longer path, compressed to the decisions that changed.
- 2013–2017
Delivery foundations
Learned to turn business constraints into maintainable enterprise and commerce software.
- 2017–2021
Commerce platforms at scale
Moved deeper into Java services, merchant systems, order domains, distributed jobs, and multi-site delivery.
- 2021–2023
Real-time systems & leadership
Worked on IoT and real-time risk platforms while raising design, review, and team-delivery standards.
- 2024–2025
Graduate engineering study
Completed graduate engineering study in Canada and broadened the systems perspective behind the work.
- 2025–Present
Platforms, products & evidence-first AI
Applying production discipline to financial platforms, an independent learning product, and developer RAG.
Field notes
Decisions, trade-offs, and what changed.
I write after doing the work: first the principle, then the implementation, and finally what the evidence changed.
RAG Is Not a Mandatory Pipeline, but an On-Demand Evidence Capability
Understand the goal and inspect the current source first; call Local RAG for evidence candidates only when cross-source information can genuinely improve the judgment.
From Workflow Harness to Lightweight Prompts: What I Removed After Codex 5.6
When a model's native workflow becomes capable enough, a Harness can shift from an efficiency multiplier to a context burden. This rewrite keeps the safety boundaries while substantially simplifying the Prompts and runtime.
Only Four Hard Rules at the Top Level: Where an Agent Must Stop
Models can become increasingly autonomous, but production, persistent data, costs, and irreversible actions must still require explicit authorization. Safety is not the opposite of efficiency; it is the prerequisite for automation.
Working through a hard systems problem?