Aland Astudillo
Biomedical engineer, data scientist and AI consultant
Gold Coast, Queensland, Australia · English and Spanish
For fourteen years I have built data and AI systems for hospitals, governments, universities and companies. Imaging pipelines, retrieval and knowledge systems, forecasting and operational reporting, and the governance work that has to happen before any of it touches a real decision. The kind that have to keep working after the pilot ends.
Selected work
- Queensland Museum and the Queensland Government
At Clevvi. Documenting First Nations cultural collections. I wrote the tender that won funding under Private Sector Pathways (PSP), $100,000 of it, then built the system over eight months with a small team: a multi-agent AI service on AWS that reads up to eight images of an artefact and drafts a catalogue record in under two minutes, with verifiable credentials for provenance through Anonyome Labs. Every record goes to curators and knowledge holders for review under ICIP governance. The system proposes, people decide. Demonstrated to more than twenty stakeholders across government, museums and industry. - A national agricultural research body
At Clevvi. I designed and built the second version of an econometric platform, extending and reworking version one: the models it ran and the system around them. - An Australian health technology startup
Through Sparkbrain. AI strategy and platform architecture, from the approach down to the shape of the system. - A Gold Coast community services organisation
Through Sparkbrain. Community data analytics and operational reporting, including the data model the reporting runs on. - Deep learning for mammography triage
Biomedical Department, Universidad de Valparaíso, in a joint pilot with the Chilean National Mammography Centre. I co-architected the deep learning cloud system designed to read mammography studies and rank them for review, so the most urgent cases could reach a radiologist first. It was evaluated in a pilot.
What I work on
I usually come in when the problem is still vague. Scoping it with the people who own it, mapping the process, designing and prototyping, then the data science itself: machine learning and statistical modelling, AI and LLM applications, retrieval augmented generation and knowledge graphs, and the AI governance and risk work that lets an organisation put any of it in front of the public.
Mostly in health, government, research and industry, where the data is messy and the constraints are real.
Experience
- Data Scientist | Senior AI and Digital Solutions Consultant, Clevvi 2025 to present
AI scoping and architecture for enterprise clients in regulated industries, document intelligence and RAG systems, and the governance frameworks that keep them auditable. - Founder and Principal Consultant, Sparkbrain 2024 to present
Independent AI, data and technology consultancy: digital transformation, process mapping, data modelling and AI applications. - Senior Data Scientist, Research Graph Foundation; Data Scientist, Swinburne University of Technology 2024
Stakeholder engagement across partner organisations, co-design and data cooperative workshops, analysis and Neo4j and Python prototypes on a national knowledge graph of the health and wellbeing research network, and the project's final report. Mentored computational science interns. - Data Scientist, then Postdoctoral Research Fellow, NICM Health Research Institute, Western Sydney University 2023 to 2025, Adjunct Fellow since
High throughput pipelines for multi channel brain signal data, statistical generative models, and noise filtering for clinical datasets. - Machine Learning Engineer, Biomedical Engineering, Universidad de Valparaíso 2020 to 2022
Deep learning and medical image analysis, moving image processing pipelines into the cloud to prioritise high risk findings for review.
Tools
Document intelligence and RAG systems, knowledge graphs, agentic systems, clinical data pipelines, time series forecasting and medical image analysis.
- Technologies Python, R, SQL, MATLAB
- AI and LLM LangChain, FastAPI, Model Context Protocol, GraphRAG, prompt engineering, evaluation
- Machine learning PyTorch, TensorFlow, scikit-learn, pandas, NumPy, statistical and generative modelling, signal processing
- Data Neo4j, knowledge graph construction and entity extraction, PostgreSQL, MongoDB, vector databases
- Cloud and delivery AWS (EC2, S3, SageMaker, AgentCore), Azure, Docker, Git, CI/CD, web and mobile development
Collaborations
- Tedix — collaborator
- Exonova — technical advisor
- Tellie — technical advisor
- Queensland University of Technology — Visiting Fellow, ARC Training Centre for Behavioural Insights for Technology Adoption, Faculty of Business and Law
What keeps my attention
Two lines of work. In industry, AI in settings where being wrong has a cost: health, government, anything with a regulator. The interesting problem there is rarely the model, it is the evaluation, the governance, and the handover to the people who have to live with it. In research, how brain activity organises itself over time, which is what my PhD was about and what I still publish on.
Applied AI writing
- Unveiling the synergy: retrieval augmented generation meets knowledge graphs — tools and platforms for integrating knowledge graphs with RAG pipelines
- The combined use of RAG and fine-tuning to improve LLM pipelines — when to retrieve, when to fine-tune, and when to do both
- What is LLMLingua? — prompt compression to improve large language model performance
- How to use GROBID to extract text from PDF files — machine-learning extraction of structured information from PDFs
Background
- BSc, MSc and PhD, Universidad de Valparaíso
Bachelor and Master in Biomedical Engineering, with a thesis on machine learning and statistical modelling in neuroscience, then a PhD in sciences, biophysics and computational biology, with a thesis on machine learning and statistical modelling applied to brain dynamics. - Certified Member , Australian Computer Society
- Registered expert , Scimex, Australian Science Media Centre
Research and writing
Peer-reviewed research in computational and cognitive neuroscience: statistical modelling of brain dynamics, signal processing and clustering, and computational models of neural activity, across both clinical and basic neuroscience. Further published work on retrieval augmented generation and knowledge graphs. Around twenty conference presentations across Australia, Chile, Canada and Italy. The full record is on Google Scholar and ORCID.
I have contributed to research programmes holding more than AUD $2M in competitive funding, including NHMRC grants APP1195709 and APP1102532, and FONDEF and FONDECYT grants in Chile.
- DeMente: el cerebro, un hueso duro de roer
“The brain, a tough nut to crack”
Popular science book on the brain, in Spanish. I wrote one chapter of it. Second edition, 2021, CINV. ISBN 9789563247213.
Profiles
Professional
Publications and research
Affiliations and expert registers
Contact
The quickest way to reach me is LinkedIn.