Bazli Nabil

Bazli standing on a George Town street under a clear sky

Solution Architect · System Design · DevOps

I design, build and scale the systems that governments and enterprises run on, and I stay accountable from the first requirement to production.

Results, measured in production

  • 40×

    Less downtime, from 96% to 99.9% uptime

  • 3.5×

    Faster record fetch, from 7s to under 2s

  • 2.4×

    Faster core workflows through performance optimisation

  • 100s–1,000s

    Concurrent users on a single platform

Where I come in

Four ways to engage, from an independent review to owning delivery end to end. For public sector agencies, enterprises and growing teams.

01

Architecture review

An independent assessment of an existing system: architecture, codebase, infrastructure and delivery practice. You receive a prioritised findings report and a practical roadmap.

Suited to planning a modernisation, validating a vendor’s design, or preparing for scale.

02

Design-to-delivery lead

One accountable owner from requirements to production. I write the specification, define the target architecture, build alongside your team and take the system through acceptance and go-live.

Suited to new platforms and the replacement of legacy or spreadsheet-based processes.

03

Scale & optimisation

For systems that are slowing down or hitting their limits as load grows. I measure first to find where time is lost, optimise the critical paths, then re-architect for horizontal scale where the return justifies it.

Suited to growing user numbers, missed performance targets and platforms that must handle more without more hardware.

04

DevOps & platform

The foundations that keep a system healthy in production: containerised delivery, CI/CD, monitoring and alerting, and centralised identity with SSO and role-based access.

Suited to teams that build well but need a dependable path to production and visibility once there.

How an engagement runs

The same shape at any size

  1. Enquiry

    You outline the system and the problem using the form below. I reply within two working days.

  2. Discovery call

    We discuss context, constraints and what success looks like.

  3. Proposal

    A written scope, approach, timeline and fee, agreed before work begins.

  4. Delivery

    Work proceeds in agreed phases, with regular demonstrations and documented decisions.

  5. Handover & support

    Documentation, knowledge transfer and an agreed period of support after go-live.

Built to scale

Scale is designed in, not bolted on. These are the decisions I make early so a system holds up at ten times the load, and the loop I run when it doesn’t.

01

Stateless services, horizontal scale

Services keep no session state of their own, so capacity grows by adding instances behind a load balancer rather than buying a bigger server. Identity sits in one place, with SSO and role-based access shared across applications.

  • Microservices
  • Containers
  • Keycloak SSO

02

Asynchronous by default

Slow or bursty work moves off the request path onto an event stream. Users get a fast response, consumers process at their own pace, and a spike in traffic queues up instead of taking the system down.

  • Kafka
  • Event-driven
  • Idempotent consumers

03

Data that keeps up

Most slow systems are slow at the database. I design schemas and indexes around the real queries, cache hot reads in memory, and batch heavy writes, so response times stay flat as records grow.

  • PostgreSQL
  • Redis
  • Query tuning

04

Observable from day one

Every service reports latency, throughput and errors from its first release. Capacity decisions and tuning are then based on production numbers, not guesses, and problems surface on a dashboard before users report them.

  • Prometheus
  • Grafana
  • Alerting

How I optimise

How 12s became under 5s

  1. Measure

    Instrument the critical paths and record a baseline under realistic load.

  2. Find the bottleneck

    Trace where time is actually spent: a query, a network hop, a lock, a payload.

  3. Fix the biggest cost

    Index, cache, batch or move work async, one change at a time.

  4. Prove it

    Re-measure against the baseline and keep the change only if the numbers move.

From 12s to under 5s

As Tech Lead and Solution Architect, I took a government valuation system from slow and unreliable to fast and stable, one measured step at a time.

01 · Load

A valuation core under load

The valuation module sat at the centre of a four-application government suite, and it could no longer keep up.

02 · Limits

Twelve seconds to save

Saving a valuation took 12 seconds and fetching a record took 7, on a system that was down about two weeks a year.

03 · Measure

Measure before changing anything

I added monitoring across the critical paths, so every tuning decision rested on production numbers rather than guesses.

04 · Re-architect

Split, cache, stream

The module became microservices, with caching for hot reads, event streaming for heavy work, and one sign-on across five applications.

05 · Proof

Measured in production

Saves dropped from 12s to under 5s, record fetches from 7s to under 2s, and uptime rose from 96% to 99.9%.

06 · Scale

Built to keep scaling

Downtime fell about fortyfold, from roughly two weeks a year to under nine hours. If your system is heading the other way, let’s talk.

Selected work

Client names are withheld. Figures are measured in production.

FlagshipPublic sectorLive

A multi-agency government platform

Solution Architect

A government programme ran a core process across roughly 80 organisations on shared Excel workbooks. It needed one system that around 1,000 people could use at the same time, delivered under MAMPU guidelines, CMMI Level 3 and ISO governance.

~1,000

Concurrent users

~80

Organisations on one system

Key decisions

  1. 01

    Model the work before replacing the spreadsheets

    Each spreadsheet process was mapped into explicit workflows, roles and validation rules before design began, so the system enforces what the workbooks only implied.

  2. 02

    Keep documents and code in step

    I led the full chain from URS through SRS and SDD to UAT, with traceability between requirements, design and implementation, so what was accepted is what was specified.

  3. 03

    Design for concurrency from the start

    Event-driven Spring Boot services with Kafka and Redis behind an Angular front end, Keycloak SSO and role-based access, containerised with Docker and observed through Prometheus and Grafana.

    • Angular
    • Spring Boot
    • Kafka
    • Redis
    • Keycloak
    • Docker
    • Grafana
Private sectorBuild

Operations system for a global manufacturer

Full-Stack Engineer

I built the Angular front end and a NestJS and PostgreSQL backend with containerised deployment, including a validated bulk-ingest endpoint and query optimisation that shortened batch processing times.

  • Angular
  • NestJS
  • PostgreSQL
  • Docker

Who you’ll work with

Bazli Nabil

I’m Bazli, a solution architect who still writes production code.

For more than seven years I have worked across government and enterprise systems, usually in the role where design decisions meet their consequences. Clients bring me in when the stakes are high and the path is unclear: a process that has outgrown its spreadsheets, a system that has to scale beyond what it was built for, or a new platform that must pass formal acceptance.

I specify, design and build the critical parts myself, document decisions as I go, and remain accountable once the system is live. The aim is always a system your team can run without me.

Disciplines
Solution architecture, system design, performance optimisation, DevOps
Sectors
Public sector, enterprise, private sector
Governance
MAMPU, CMMI Level 3, ISO
Response
Within two working days

Tell me about the system

A few details are enough to begin. I will reply with initial questions and, where it is a good fit, propose a discovery call.

Everything you share is treated as confidential. I reply within two working days.

Or email bazli.rekara@gmail.com