Kecia de Moura

At NeuroTech Information Technology, I built and maintained data ingestion pipelines and integrations that fed downstream analytics and BI systems, working with REST and Google APIs to bring external data into internal platforms. More recently, I've explored Databricks Lakeflow Connect for batch and streaming ingestion patterns, including change data capture and JSON flattening.

Across roles at NeuroTech, Pivot Subscriptions, and House Hound, I've worked hands-on with relational databases (MySQL, SQL Server, PostgreSQL), NoSQL storage (DynamoDB), object storage (AWS S3), search infrastructure (Elasticsearch), and GraphQL-over-Postgres tooling (Hasura). I've also recently completed Databricks Academy training on Delta Lake fundamentals.

My PhD research required building end-to-end data processing pipelines: from raw signature images through feature extraction, data-driven summarization, and structured training-set generation, all reproducible and version-controlled in Python. Earlier in my career, I processed and transformed data professionally using SQL and R at NeuroTech. I've since begun exploring distributed processing with Spark via Databricks Academy coursework.

As a Senior Data Analyst at NeuroTech (later acquired by B3), I designed end-to-end BI solutions: modeling star-schema data warehouses, building OLAP cubes with Mondrian via Pentaho Schema Workbench, writing MDX queries against those cubes, and building HTML-based dashboards and ad hoc reports served through the Pentaho BI Server. I also automated monitoring dashboards to track application performance and support data-driven decision-making across the organization.

This is the core of my PhD research at ÉTS Montréal. I've designed novel training-data generation strategies for biometric verification systems (ProtoSig), built adaptive stream-based learning systems for signature verification, and conducted rigorous statistical validation (frequentist and Bayesian) to demonstrate the reliability of my methods. My work received the Best Paper Award at ICPR 2024, and I've published in Pattern Recognition and Pattern Recognition Letters.