Nairobi, Kenya · Building across backend, native mobile and infrastructure

Senior software engineer

I build the systems people depend on.

From real-time communication to transaction-safe financial infrastructure, I turn difficult product problems into calm, observable, production-ready systems.

See the evidence Download CV

01 / Position

Seniority is not how much code you can produce. It is knowing where complexity belongs, what can fail, and how the team will know.

02 / Selected system work

Evidence, not a highlight reel.

One growing platform. Several difficult systems. Each case note focuses on the decisions beneath the interface.

Platform architecture2023 — now

WasaaChat: building a platform, not a pile of features.

Architecture and hands-on implementation across identity, messaging, calling, feeds, commerce, media, notifications, native clients, administration and release operations.

75kusers onboarded
Multi-clientiOS, Android and web
Domain-ledclear service boundaries

Core platform

Identity

Realtime

Money

Media

Realtime systems02

Communication that survives the real world.

Presence, delivery state, offline queues, background synchronization, multi-instance fan-out and SFU-backed audio/video calling.

Financial infrastructure03

Money systems designed for failure.

Exact values, append-only double-entry records, idempotent operations, callbacks, reversals, risk checks and reconciliation.

03 / How I think

A system is a set of promises.

Select a layer. The interface changes; the engineering obligation does not.

Selected principle

Put complexity where it can be owned.

Good boundaries make change local. Domains should expose stable contracts while keeping their invariants, data and failure policy explicit.

domain ownershipstable contractsmigration safety

04 / Operating method

Calm systems come from deliberate work.

01
Model the truth
Start with invariants, actors and state—not endpoints or screens.
02
Design the failure
Retries, stale data and partial completion are normal operating conditions.
03
Instrument the claim
If a system says it is healthy, there should be evidence behind the statement.
04
Release reversibly
Safe migrations, staged delivery and honest fallbacks keep change survivable.

05 / Field notes

What the incident teaches.

Short notes from building systems where the difficult part begins after the happy path.

A timeout is not a verdict.

A response can disappear after the operation commits. Financial and distributed workflows therefore need stable idempotency keys, explicit state and reconciliation—not blind retries.

Transactions · failure semantics

The client lifecycle is part of the system.

Background execution, permissions, network transitions and operating-system delivery rules are distributed state. Treating them as interface details creates failures no backend dashboard can explain.

Native mobile · realtime

Observability is a product feature.

Useful diagnostics preserve context: the operation, participant, role, transition and correlation ID. That evidence turns a two-device mystery into a sequence the team can reproduce and fix.

Operations · engineering evidence

06 / About

Breadth without hand-waving.

I have grown by owning complete outcomes: the service, the client, the migration, the failure path and the operational follow-through. That is how I learned to see products as systems rather than collections of tickets.

What I bring

Backend and distributed architecture
Native iOS and Android engineering
Transaction integrity and financial workflows
Security, identity and administrative controls
Production diagnosis and release readiness

Tools I reach for

TypeScript · Go · Swift · Kotlin · SQL · PostgreSQL · Redis · Kafka · WebSockets · SwiftUI · Jetpack Compose · Docker · Kubernetes · OpenTelemetry

Beyond delivery

I lead technical planning and code review, communicate across product and operations, and mentor final-year IT students at Strathmore University.

07 / Next

Have a difficult system to make simple?

I’m interested in ambitious engineering problems, responsible AI systems, and teams that value evidence over theatre.

salvatoluice5@gmail.com