AdamJarick.

Builds software for the Building Commission at NSW DCS.

↓Scroll to enter

✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney✦Engineer✦AI Systems✦Full-Stack✦Cloud✦Distributed✦NSW DCS✦Sydney

001 / About

Hey. I'm Adam, a software engineer based in Sydney.

I'm Adam, a Software Engineer based in Sydney with a double degree in Software Engineering (Honours) and Cyber Security from Macquarie University. I build things at the intersection of AI, full-stack development, and cloud infrastructure.

Currently at the NSW Department of Customer Service, I'm building software solutions for the Building Commission. Previously at Apate.AI I architected and deployed anti-scam systems processing thousands of messages daily across WhatsApp, Telegram, and email, and before that I built voice authentication neural networks at Webschool.au.

When I'm not shipping for the Building Commission, I'm training poker AI to play Nash equilibrium strategies or tinkering with whatever catches my interest next.

002 / Track Record

The track record.

Apr 2026 → Now

Software Engineer

NSW Department of Customer Service · Sydney, NSW

◦Own the entity resolution platform that consolidates hundreds of millions of names and companies from close to 100 source systems into deduplicated, mutually enriched master tables, built on Splink probabilistic record linkage in Databricks.
◦Built and own the Elasticsearch layer that gives analysts and building inspectors fuzzy name and company search across those hundreds of millions of rows, plus the risk rating that directs inspectors towards the sites most likely to be non-compliant.
◦Sole full stack engineer on a team of 15+ otherwise specialised in backend and data engineering, running ETL in Azure Data Factory and Databricks and owning the Postal/SMTP alerting the analytics team and field inspectors rely on.

100M+records linked

~100source systems

Splinkentity resolution

PythonTypeScriptSQLAzure DatabricksAzure Data FactoryMS SQL Server

Oct 2024 → Apr 2026

Software Engineer

Apate.AI · Town Hall, NSW

◦Joined as one of the first engineers and helped grow the platform from zero to over $1M in annual revenue.
◦Sole architect and developer of the anti-scam conversational agents for WhatsApp, Telegram and email, built end to end in TypeScript with WebSocket manipulation and OpenAI models to engage scammers.
◦Cut voice bot response latency from 1.2 seconds to 150 milliseconds by standing up an Australian ElevenLabs endpoint on Cloudflare edge infrastructure.

$1M+annual revenue

150msvoice latency

100K+intel artefacts

TypeScriptReactNode.jsOpenAIDockerKubernetes

Jun 2025 → Nov 2025

AI Engineer Intern

Webschool.au · Macquarie, NSW

◦Architected and built a Voice Authentication Neural Network for a platform with the purpose of two-factor authentication, breaking down business objectives into actionable tasks using Agile methodologies.
◦Utilised the PyTorch library to create a Siamese Neural Network which compared two sets of voices to authenticate if they were the same voice for login purposes, using the VoxCeleb database for training with an input of 13 Mel-frequency cepstral coefficients per audio frame and cosine similarity.
◦Deployed on the company's portal, interweaving its functionality with existing infrastructure.

PythonPyTorchAzureDocker

2021 → 2025

BE (Hons) Software Eng · Cyber Security

Macquarie University · Sydney

◦Double degree in Software Engineering (BE Honours) and Cyber Security (Bachelor of Information Technology), graduating with a Distinction average.
◦Honours thesis: AI Co-Pilot, a Chrome extension delivering personalised learning recommendations from quiz performance and engagement signals.
◦Built a Deep CFR poker agent on the side that converges toward Nash equilibrium.

AlgorithmsCryptographyNetworksMLDistributed Systems

003 / Toolkit

The toolkit.

001

PythonLanguages

002

TypeScriptLanguages

003

JavaScriptLanguages

004

GoLanguages

005

SQLLanguages

006

HTMLLanguages

007

CSSLanguages

008

ReactFrontend

009

Next.jsFrontend

010

AngularFrontend

011

Vue.jsFrontend

012

FlutterFrontend

013

TailwindFrontend

014

Three.jsFrontend

015

ElectronFrontend

016

Node.jsBackend

017

ExpressBackend

018

DjangoBackend

019

FlaskBackend

020

BunBackend

021

REST APIsBackend

022

JWTBackend

023

WebSocketsBackend

024

PostalBackend

025

KeycloakBackend

026

PostgreSQLData

027

MS SQL ServerData

028

ElasticsearchData

029

SplinkData

030

SQLiteData

031

MongoDBData

032

DynamoDBData

033

DatabricksData

034

i2 iBaseData

035

PrismaData

036

RedisData

037

ClickHouseData

038

PyTorchData

039

BurnData

040

ValkeyData

041

DockerDevOps

042

KubernetesDevOps

043

AWSDevOps

044

AzureDevOps

045

TerraformDevOps

046

GitHub ActionsDevOps

047

HelmDevOps

048

GrafanaDevOps

049

CloudflareDevOps

050

JestDevOps

051

PlaywrightDevOps

004 / The Work

The work.

PRJ/01 · OPENGTO · LIVEPOT 4.0 BB

You are BTN · ButtonAKo · vs 2.5BB open

Pot

4.0

UTG

A♠

K♥

Hand 1 / 7

PRJ / 01

OpenGTO.

Neural network-powered poker preflop trainer converging toward Nash equilibrium.

Deep Counterfactual Regret Minimisation (Deep CFR) with three networks self-playing across 20K+ training iterations. 317-D feature vector, 8-stage curriculum learning, weighted reservoir sampling over 5M+ samples. Cross-platform Electron desktop app with interactive table, live GTO feedback, and a Range Viewer for all 169 starting hands.

20K+training iters
5M+samples
317-Dfeature space
169hand ranges

PyTorchDeep CFRReactElectronFlaskONNX

Open full trainer GitHub

CSA Co-Pilot

Hello!
Welcome to the CSA Co-pilot.

Please login using your student ID to get started.

Enter your Student ID

PRJ / 02

AICo-Pilot.

Chrome extension turning learning analytics into personalised recommendations.

Honours thesis project. Reads quiz scores, engagement time, skipped lessons and link clicks from SCORM and Moodle, then surfaces personalised LLM-driven recommendations through a Chrome extension. Demonstrated lift in student engagement and module completion.

Moodle + SCORMLMS platforms
+35%engagement
Honoursthesis
2person team

JavaScriptChrome APIsSCORMMoodleLLMREST APIs

View deep-dive GitHub

005 / Dossier

The full record.

Every role, project and number in one place. Two pages, kept current.

Download résumé

View in browser

PDF

2 pages

Updated Sept 2026

On the page

Full stack · AI/ML

Building Commission NSW

Bugcrowd P1 · CVSS 9.8

Baseline clearance

006 / Ground

AdamJarick.

Hey. I'm Adam, a software engineer based in Sydney.

The track record.

Software Engineer

Software Engineer

AI Engineer Intern

BE (Hons) Software Eng · Cyber Security

The toolkit.

The work.

OpenGTO.

AICo-Pilot.

The full record.

Let's build something.