AI & Data Engineer

Building AI-Powered Solutions

AI & Data Engineer with 5+ years of experience designing and implementing scalable data pipelines, cloud infrastructure, software solutions and intelligent automation, along with 2+ years focused on AI systems.

+0
Years Experience
0
Cloud Certifications
Let's talk Download CV

// services & products

Services

What I can build for you, by category. Each category includes examples of work I have already delivered, with live demos where available.

Custom AI Systems ERP & Operations Platforms Web Apps & Landing Pages Data Platforms & Cloud Migration AI Agents & n8n Automations RAG & LLM Assistants
01

Custom AI Systems

End-to-end AI products: from the model and the data pipeline to the interface. OCR extraction, AI coaches, conversational analytics, vision.

What you get

  • Document OCR + LLM extraction
  • AI coach & conversational products
  • Dashboards that answer questions
Python Azure ML Databricks Dash/Plotly
Zenda — AI Financial Coach preview
Live site

Zenda — AI Financial Coach

Workplace-benefit app that turns what's left after the bills into a dated roadmap for one real goal: a trip, a car, a home. One payday check-in, and an AI coach running on Claude that explains every number using the user's own data, never guesses. Built and validated at Weekend of Startups with Swyftx, UQIES and UQ Ventures.

Claude AI Coach Fintech Product Design Startup

Computer Vision Tracking

Computer vision pipeline built with neural networks models such as YOLO and SAM2 to track animals movement for Queensland Brain Institute.

Python YOLOv8 SAM2 Computer Vision
02

ERP & Operations Platforms

Operations systems designed from scratch or configured to your workflow: inventory, orders, dispatch, roles and live dashboards in one place.

What you get

  • Custom-built ERP modules
  • Configuration of existing tools
  • Centralised data + reporting
PostgreSQL Node.js Streamlit Power BI
03

Web Apps & Landing Pages

Fast, modern websites and web apps with AI features built in. Designed, developed and deployed, with analytics from day one.

What you get

  • Landing pages & product sites
  • Dashboards and internal tools
  • Deployed on Vercel or Azure
Astro React Tailwind Vercel
IDEA Bio preview
Visit site

IDEA Bio

Website for IDEA Bio, an NCRIS and UQ-funded biofoundry offering synthetic biology services to Australian researchers and businesses: services, a look inside the facility, partners and contact.

Corporate Website Research SEO
GB Safety Solutions preview
Visit site

GB Safety Solutions

Bilingual website for an occupational health and safety consultancy: services, management systems, audits, training, team and contact.

Business Website Bilingual Vercel
Miluska Buelot — Portfolio preview
Visit site

Miluska Buelot — Portfolio

Professional portfolio for an industrial engineer and occupational health and safety consultant: profile, experience, services and contact.

Portfolio Landing Page Vercel
04

Data Platforms & Cloud Migration

Pipelines, lakehouses and migrations that hold up in production: from legacy mainframes and SQL Server to Azure, Databricks and Snowflake.

What you get

  • Azure Data Factory + Medallion lakehouse
  • Spark batch pipelines & validation
  • Power BI analytics layer
Azure Data Factory Databricks Spark Snowflake
Predictive Maintenance Lakehouse preview
Live Demo

Predictive Maintenance Lakehouse

Predictive Maintenance Pipeline for Industrial IoT using lakehouse architecture.

Jupyter IoT Machine Learning Data Pipeline
Traffic Analysis Spark HDFS preview
Live Demo

Traffic Analysis Spark HDFS

Big data pipeline built with Spark using HDFS files to analyze traffic and identify causes of it in Sydney.

Spark HDFS Docker Big Data
05

AI Agents & n8n Automations

Agentic workflows that run your operations: WhatsApp, email, ERPs, databases and dashboards orchestrated end to end with n8n and LLMs.

What you get

  • Incident intake & classification bots
  • ERP ↔ dashboard ↔ WhatsApp sync
  • Approval flows, alerts and reports
n8n LLMs WhatsApp API Webhooks
MineOps Dispatch preview
Demo based on request

MineOps Dispatch

Agentic AI system that automates mining field operations via WhatsApp. An AI agent receives operator reports, classifies incidents, updates the control room database, and sends real-time confirmations — fully autonomous, end to end.

n8n AI Agent WhatsApp Vercel
06

RAG & LLM Assistants

Assistants that answer with your own documents and data, never guesses. Retrieval pipelines, evaluation and full query traceability included.

What you get

  • Knowledge-base chat over your docs
  • Code & SQL migration copilots
  • Semantic search on Azure AI Search
RAG Azure AI Search HuggingFace Claude
Databricks RAG Migration Tool preview
Live Demo

Databricks RAG Migration Tool

AI-powered system that transforms legacy T-SQL into modern Databricks SQL using RAG, LLM reasoning, and deterministic validators for safe, reliable migration.

Python RAG LLM Databricks Azure Azure Foundry

// process

How I work

01

Discovery

A short call to understand your problem, your data and what 'done' looks like.

02

Prototype

A working version in days so you can test it with real inputs before committing.

03

Production

Hardened, deployed, documented and handed over. Support available after launch.

Tell me what you need

Available for freelance projects.

// about

About me

Catherine Varas
RAG & Agentic AI
Azure ☁
Data · AI · Cloud

I'm Catherine, an AI & Data Engineer with more than 5 years of experience. I design and build data pipelines, cloud platforms and AI systems, and I've done it for a bank, a mining company, a research institute, consultancies and startups. I work across the whole stack: data, backend, AI models and the interfaces people use.

Some of what I've built: moving a bank's core records from legacy systems to the cloud; bringing a mining company's operational data into one place for reporting and analytics; AI tools that automate migrations and accounting work for consulting clients; automations and an AI assistant that run daily operations for teams over WhatsApp; and an analysis and simulation platform for a research lab. In every case I learn the business, understand the data and deliver something that works.

  • End to end

    Data, backend, AI and the interface your users touch. One accountable owner.

  • Production-grade

    Tests, CI/CD, validation and documentation come standard, not as extras.

  • Fast to value

    A working prototype in days, then we iterate to production together.

Available for freelance projects.

See services Get in touch

// tech stack

Tech stack

Software Engineering

Python Python TypeScript JavaScript Java Java SQL R C# Bash Node.js Node.js React Astro Docker Docker GraphQL GraphQL REST APIs

Cloud

Azure Azure Azure ML Azure ML Azure Functions Azure Functions Azure AI Search Azure AI Search Azure Data Factory Azure Data Factory AWS AWS GCP HPC GPU Streaming

AI & ML

YOLO SAM2 Neural Networks HuggingFace RAG scikit-learn scikit-learn Deep Learning Computer Vision Agentic AI LangChain n8n

Tools

Dash/Plotly Power BI Power BI Streamlit Streamlit matplotlib Git Git CI/CD

Databases

Oracle Oracle SQL Server PostgreSQL PostgreSQL MySQL MySQL Elasticsearch

// experience

Experience

Click a role to see the details.

  • ▹ Integrated data across multiple mining domains and system types, including operational databases, equipment and telemetry systems, production systems, and other enterprise sources, enabling a unified view of mining operations through designed and developed Azure Data Factory pipelines that integrate and consolidate data into a centralised data lake for Power BI analytics and machine learning applications.
  • ▹ Implemented a Medallion Architecture (Bronze, Silver, Gold) to progressively ingest, standardise, validate, and transform raw operational data into trusted analytical datasets, creating a scalable foundation for enterprise reporting, advanced analytics, and machine learning.
  • ▹ Designed and built an agentic automation platform from the ground up, integrating n8n and LLMs to orchestrate end-to-end business processes and synchronise ERP systems, dashboards, and real-time WhatsApp communication for operational teams.
  • ▹ Developed a custom operations management platform from the ground up, combining ERP functionality, workflow automation, centralised data management, and interactive dashboards into a unified system tailored to business operations.
  • ▹ Built an AI-powered analytics platform from the ground up, featuring a conversational interface that answers data questions and dynamically reconfigures visualisations through natural-language requests, leveraging RAG and LLMs with full end-to-end query traceability.
  • ▹ Developed an end-to-end data analysis and simulation system for bio-experiments, integrating heterogeneous bioreactor, sample, and metabolomics data and transforming raw experimental data into validated metrics, visual insights, and actionable results across multiple bioreactor platforms.
  • ▹ Built an interactive environment for analysing and simulating fermentation processes, allowing researchers to explore experiments, modify parameters and variables, recalculate results dynamically, compare runs, and test different fermentation scenarios within a single workflow.
  • ▹ Engineered the platform using an AI-assisted development workflow with Claude Code, combining Sonnet and Opus agents with modular architecture, automated testing, data validation, reproducible workflows, and iterative code review — delivering a sophisticated research tool designed for real-world bioprocessing and presented at GBA 2026 in France.
  • ▹ Built a multi-model pipeline (YOLO + SAM2) for automated animal tracking
  • ▹ Engineered chunked GPU streaming and HPC parallel execution for long video processing
  • ▹ Designed a geometric classifier detecting 6 social interaction types from spatial coordinates
  • ▹ Reconstructed neural network architectures from raw checkpoint weights
  • ▹ Architected and engineered cloud-based AI systems on Azure, integrating Azure Machine Learning, Durable Functions, Azure AI Search, Key Vault and Azure DevOps CI/CD to deliver scalable AI orchestration, HuggingFace model serving, Logic App integrations, and cost-optimised production workflows.
  • ▹ Developed an AI-powered migration platform that combines metadata-driven processing and Retrieval-Augmented Generation (RAG) to automate and accelerate enterprise analytics migrations, including QlikView → Power BI and SQL Server → Snowflake, reducing manual mapping and improving migration consistency and accuracy.
  • ▹ Engineered an AI-driven accounting automation system on Databricks, combining OCR, structured signal normalisation and LLM-based semantic resolution to transform unstructured financial data into analysis-ready datasets and feed downstream calculation engines through automated Databricks Jobs and deployment pipelines.
  • ▹ Built interactive analytics and migration intelligence applications using Streamlit and Python, providing dependency mapping, migration-gap detection, automated scoring, validation, and data-quality assessment across complex analytics environments.
  • ▹ Led migration of millions of contractual records from IBM Mainframe to cloud infrastructure using Apache Spark
  • ▹ Designed frameworks for data quality, scalability, and compliance with banking regulatory standards
  • ▹ Developed Java backend routines dynamically querying Mainframe or Oracle systems
  • ▹ Managed centralized contracts module exposing data across banking systems
  • ▹ Engineered scalable batch pipelines using Apache Spark and Java for multi-product Mainframe-to-cloud migration
  • ▹ Developed reusable Spark routines for data transformation, validation, and reconciliation in Parquet format
  • ▹ Automated business-rule validations and anomaly detection across millions of contract records
  • ▹ Deployed and maintained 70+ reusable components (Java backend, Spark batch, database scripts)
  • ▹ Built Big Data pipelines for performance reporting; automated daily processing of 200+ SFTP files
  • ▹ Designed dashboards for inventory and sales forecasting for multinational retail clients
  • ▹ Built integration pipelines between Hasura GraphQL and SQL Server

// certifications

Credentials

Microsoft

Fabric Data Engineer Associate

DP-700

Microsoft

Microsoft

Fabric Analytics Engineer Associate

DP-600

Microsoft

Microsoft

Power BI Data Analyst Associate

PL-300

Microsoft

AWS

Cloud Practitioner

CLF-C02

AWS

Microsoft

Azure Fundamentals

AZ-900

Microsoft

Databricks

Generative AI Fundamentals

Databricks

// education

Academic Path

University of Queensland

Master in Data Science

2025 – Jun 2026

Brisbane, Australia

Cloud Computing Machine Learning AI Deep Learning Bioinformatics

University of La Rioja (UNIR)

Master of Artificial Intelligence

2024 – 2025

Spain

Generative AI Machine Learning Computer Vision Neural Networks

Peruvian University of Applied Sciences (UPC)

Bachelor of Industrial Engineering

2019 – 2023

Lima, Peru

Project Management Process Optimization

Institute Cibertec

Associate's in Computer Science

2019 – 2022

Lima, Peru

SQL Java Python Web Development

// contact

Let's make your idea real

Have a process to automate, a product to build or data you want to put to work? Tell me what you have in mind and I'll help you turn it into a working system.

Get in Touch

Download my CV (PDF)