Andy Challis

Assessed and estimated migration effort for 50+ applications moving from on-premise Postgres/Oracle to AWS Redshift.

Developed a custom AST parsing tool to extract key SQL migration metrics and automate effort estimation.

Mapped dependencies across Python/R/Node.js applications, identifying blockers and migration risks.

Designed a strategy to transition Rundeck jobs to Airflow and refactor raw SQL scripts into dbt workflows.

Provided actionable insights to optimize SQL translation for Redshift compliance while ensuring performance.

Developed a GraphRAG-based AI system combining knowledge graphs (Neo4j) and vector databases (pg_vector) to enhance factual accuracy.

Implemented multi-LLM token-based cost optimization, dynamically routing requests to different models based on budget and complexity.

Built custom LangGraph agents capable of generating and executing Cypher queries for structured knowledge retrieval.

Deployed a scalable fullstack AI architecture using SST (Pulumi), Next.js, AWS ECS, RDS, and ECR.

Accelerated GenAI adoption at Slalom, with the solution reused in multiple enterprise AI projects.

Lead the technical architecture of the microservices, advising on software architecture, feature delivery and testing strategy

Created multiple PoC's to test the feasibility of the project, including a 'proof of concept' for the integration of Kafka, Commercetools and Algolia

Created a backlog of work for the team to deliver on with a roadmap and t-shirt sized estimates for the work to be done

Lead the technical delivery of the microservice, advising on software architecture, feature delivery and testing strategy

Designed custom data model from scratch to emulate Google Drive including nested folders, sharing permissions and file versioning

Collaboratively worked with SafetyCulture to release the feature to end users and gather feedback for future iterations

Created a sophisticated plugin system that allowed developers to inherit from a base class which would make use of the type hints to automatically create the frontend form input for configuration

Pioneered the use of distributed tracing within the Prezzee teams, allowing for the existing rich monitoring dashboards to continue to be supported in the microservice world

Deployed using containerization on AWS Fargate with auto scaling, running on minimal hardware which shrunk the AWS costs considerably down to 100's per month for a globally distributed service

Supported the next phase of work for managing Inventory and integrating with the new Merchants microservice

Migrated existing HTML Jinja templates to chosen SaaS provider called Stripo which allowed for a non developer to modify the templates in a drag-and-drop manner

Migrated Datadog dashboards to the new microservice to ensure that the team could monitor the new system in the same way as the old system

Lead the technical delivery of the application, advising on software architecture, feature delivery and testing strategy

Collaboratively worked on features/bugs in both the backend and the frontend of the application

Stretched laterally across into Software Engineering to support the delivery on an upcoming deadline

Supported the team as a technical advisor and collaboratively sculpted the solution design document

Utilized previous open-source work and development accelerators to build application scaffolding to ensure that team members could hit the ground running and not be blocked for their build tasks

Worked with internal stakeholders, both technical and functional to ensure solution met all requirements for data movement throughout multiple environments

Advised the team on the technical direction to move around blockers in innovative ways making use of fake data and reducing coupling points

Proactively groomed and managed the backlog of work to ensure sprint planning sessions were short and sharp to reduce meeting effort required by team so they could concentrate on the problems at hand

Researched and architected a platform capable of handling customer, advertising and streaming data across multiple channels (digital and linear)

Collaboratively crafted a high level technical implementation roadmap with top down approximations for both team size and skill estimation and overall costings both CapEx and OpEx

Presented and reviewed the strategy with senior stakeholders from FoxTel, incorporating feedback into both the roadmap and architecture

Developed and deployed an event driven solution to production with the source code developed externally to SvA, showcasing the new preferred delivery model by separating the need for data from the solution

Designed an open-source JSON schema data faker that enabled the delivery of this project to simulate the data that would be present in production. The open source project is available here - jsf

Solution Lead for a $5 million-dollar account

Provided technical direction for a fully remote team of 35 people, plus client stakeholders

Deployed a production deep learning system in centerlink with an error rate of < 0.1% serving hundreds of thousands of Australians per day

Managed the delivery of 6 products, 100% on time and beyond the client’s expectations, adapted to the client’s requests and changing stake holder landscape by rapidly iterating on feedback and improvements showing the flexibility and adaptability of our team

Organized weekly lunch and learns, pair programming sessions, and out of hours mentorship

Considered a role model project, both internally and on the client side during corona virus pandemic, demonstrating how the remote working model can scale

Architected and built a self service micro-service for registering data sources to a data platform specifically for files being sent via SFTP.

Built a horizontally scalable (EC2/Docker) Python daemon that monitored file-system events (EFS) and triggered custom data flows for each object, such as; moving to specified dynamic S3 paths (S3), triggering a custom transformation job (Jenkins), automated un-compressing, clearing from file system (EFS)

Advised team members on deployment strategies (Terraform), unit testing (PyTest), alerting/logging (SNS + Splunk), documentation standards and containerization

Lead an automation squad consisting of 8 people within NAB to help deliver a centralized self service data platform.

Had a major impact on software practices within squads by; creating reusable distributed packages for other teams to take advantage of, building tactical tools to increase the work-flow and automation of manual data ingestion – effectively cutting ingestion time from 6 weeks down to 2 (allowing for business approvals and procedures).

Built user interfaces (Flask, Javascript, HTML, CSS) for other squads to use that speed up their validation efforts for data ingestion.

Created a private light GUI client which would be used to verify and process the output of the models predictions by an agent.

At the end of the trial the result was over 25 times faster throughput of forms; which in turn saved HMRC staffing from peak numbers of 120 down to 10. The output of the validation also created training data to be used to create a more sophisticated model in the future.

Created unit tests and a full web front-end as well as a mobile application developed in Swift.

Developed a custom suite of Python libraries to speed up development, such as connecting to our JIRA instance, logging/interacting with Rocketchat and connecting to our database & incorporating commonly used functions/data types.

Worked with the platform team to push Continuous Integration with GitLab using Jenkins and Docker containers.

Technical lead on a scenario to engineer full-scale automated pipelines from inception through to beta phase, presented to users in a front-end tool.

Developed algorithms for detecting leaks in pipes using multiple data sources: pressure, flow, pipe attributes, environment and smart meters.

Mentoring junior staff and delivering `lunch and learn' talks on hot topics.

Created tutorials for ways in which to achieve an out-of-the-box baseline result using TensorFlow.

Curated images and videos for the competition from both paid-for and CC0 sources.