Pipelines · Schemas · Scale
Five years building schemas, ingestion pipelines and transformation systems — the layer that turns encrypted flat files, field surveys and GIS exports into something a reporting stack can trust. Alden Systems now, Medallia before that. Maths degree, which mostly shows up as refusing to ship a pipeline I can't reason about.
$ cat profile.json { "role": "Data Engineer", "focus": ["schemas", "ingestion", "transformation"], "scale": { "records": 4_000_000, "hierarchy": 166_000 }, "stack": ["Python", "SQL", "MongoDB", "SFTP"], "degree": "B.S. Mathematics", "employer": "Alden Systems", "since": 2021 } $ _
01 — How I build
Start at the schema. How a record is parsed, structured, stored and read downstream is the contract. Get it wrong and every query after it is a workaround — so the schema is designed before the pipeline, not discovered by it.
Automate the boring, dangerous parts. Fetching, decrypting and parsing files by hand is where data quietly goes missing. If a step happens weekly, it belongs in code — fetch → decrypt → parse → load, unattended.
Reconcile, don't assume. A pipeline that runs without error is not the same as a pipeline that moved everything. Counts in, counts out, and a way to prove it.
Profile before optimizing. The bottleneck is rarely where it feels like it is. Measure, then fix the complexity — that's how -30% time complexity and -23% load time happened, rather than by guessing.
Translate, don't relay. Utilities and product owners describe outcomes, not constraints. Turning "the pole owner needs this" into something a pipeline can actually enforce is most of the job — and skipping it is how systems get built twice.
// ingestion path built at Medallia — unattended, scheduled, reconciled
02 — Selected work
Architected the document schemas and data importers governing how millions of incoming survey records were parsed, structured, stored and accessed by every downstream analytics workflow. The schema was the contract — get it right and the analytics layer stops being a guessing game.
Designed ingestion workflows that fetch, decrypt and parse flat files from enterprise sources straight into MongoDB collections — removing a manual, error-prone handoff and turning a recurring chore into infrastructure nobody has to think about.
Orchestrated upstream ingestion from client SFTP feeds, processing 166,000+ employee records into organizational hierarchy structures that client-facing reporting systems depend on — then built the transformation pipelines integrating hierarchy and account data for scoped feedback analysis and action planning.
Administered and optimized the data structures and algorithms behind core systems, cutting time complexity by 30%, and reduced system load time by 23% using efficient cloud-based technologies. The kind of work where the win is measured in what users stop noticing.
Engineering and synthesizing large-scale asset datasets from multiple sources — turning raw field and GIS data into clean, structured outputs for utility asset management. Translating requirements from pole owners and power utilities into workflows, and owning continuous improvement across the AldenOne transformation pipelines through to UAT signoff.
03 — Stack
04 — Path
Product Specialist / Data Engineer · Alden SystemsCurrent
Birmingham, AL — field and GIS asset data, AldenOne transformation pipelines, cross-functional delivery through UAT.
Senior Technical Analyst · Medallia
Birmingham, AL — document schemas and importers at 4M+ record scale, automated SFTP ingestion into MongoDB, mentoring junior analysts on architecture and implementation.
Technical Analyst · Medallia
Birmingham, AL — org-hierarchy construction from client SFTP feeds, transformation pipelines into reporting, algorithmic and load-time optimization.
B.S. Mathematics · Birmingham-Southern College
Aug 2017 – May 2021.
05 — Contact
Open to data engineering, pipeline and platform work — and happy to talk schemas with anyone who has strong opinions about them.