Data engineering
& information design.
Applications that make data useful.
Amicus Data is a data engineering operation — analytics platforms, interactive dashboards, and data products for professional services industries. Everything from raw data ingestion to the final visualization, all custom-built from scratch.
What we actually do.
Raw data — messy, scattered, high-volume, arriving from dozens of sources — gets turned into something a business can actually use. Not a spreadsheet. Not a static report. A living system that updates itself, catches what matters, and presents it in a way that makes the next decision obvious.
That means the full pipeline: an ingestion layer that pulls data in, transformation logic that cleans and structures it, a warehouse that stores it efficiently, and a frontend that makes it legible and interactive. Every layer built in-house. No outsourcing, no drag-and-drop tools, no templates.
The platforms underneath — Google Cloud, BigQuery, cloud compute — provide essentially unlimited storage and processing power. That's the raw infrastructure. Everything built on top of it is entirely custom: the extraction logic, the classification models, the orchestration, the APIs, the dashboards, the data products.
Nothing here is resold or reconfigured off-the-shelf software. The code does exactly what the problem requires, and nothing more. That's a fundamentally different approach from plugging into an existing BI tool, and it produces fundamentally different results.
Data Engineering & ETL
Extraction, transformation, loading — the heavy infrastructure work. Pulling data from APIs, web crawls, databases, and third-party platforms. Cleaning it, normalizing it, structuring it. Moving it reliably at volume into warehouses where it becomes queryable and useful.
Systems Architecture
The interconnect — how different systems talk to each other and what happens in between. The intelligence layer that sits on top of the raw infrastructure, orchestrating data flows, managing dependencies, and making sure every piece of the system knows what it needs to know, when it needs to know it.
Information Design
The part most data projects skip. How information is organized, structured, and presented so it actually communicates. The layouts, the visual hierarchies, the content architecture — the design decisions that determine whether data gets understood or just gets ignored.
Dashboards & Visualization
Interactive, numbers-heavy frontends with real depth. Charts, graphs, heatmaps, ranking tables, competitive grids — dense with data but designed to be navigable. The kind of interfaces where you can spend an hour exploring and keep finding things you didn't expect.
How we build.
Every project starts the same way: data exists somewhere — usually in multiple somewheres — and it needs to be somewhere else, in a different shape, doing something useful. Extraction is the first problem. APIs have rate limits, websites have anti-scraping measures, databases have schemas that were designed in 2009. Getting the data out cleanly and reliably is its own discipline.
Transformation is where the real work happens. Raw data is inconsistent, duplicated, and full of edge cases. The transformation layer is where business logic lives — the rules that turn a messy API response into a structured, queryable record that means something specific. That logic gets written by hand because the edge cases matter, and generic tools don't know about anyone's edge cases.
Loading is the part that sounds simple but isn't. Getting data into a warehouse efficiently — incrementally, without duplication, with proper partitioning and schema evolution — is a problem that scales in complexity with volume. At 660 million records a month, the loading strategy matters as much as the transformation logic.
The most interesting part is the systems layer — the interconnect. Individual pipelines are useful. A system of pipelines that are aware of each other, that share state, that can trigger downstream processes when upstream data changes — that's where things get powerful. The intelligence layer sits on top of the raw infrastructure and orchestrates everything: scheduling, dependency management, error recovery, monitoring.
The cloud platforms — Google Cloud, BigQuery — provide effectively unlimited compute and storage. Petabyte-scale data warehousing, elastic processing, global infrastructure. That's the raw material. Everything built on top of it is entirely custom: the extraction logic, the transformation rules, the orchestration, the APIs, the publishing layer, the dashboards. All of it written from scratch.
The raw platforms provide immense compute and storage capacity. Everything else — every line of application logic, every interface, every orchestration rule — is custom software. That's a fundamentally different approach from configuring off-the-shelf tools, and it opens up capabilities that template solutions simply can't reach.
"The data problems, the engineering puzzles, the moment a dashboard makes something visible that wasn't visible before — that never gets old."
I'm Douglas Mallett.
Data engineer, analyst, and developer. Amicus Data came out of a problem that kept coming up: interesting data trapped in systems that couldn't do anything useful with it. Raw APIs returning millions of records with no structure. Databases full of valuable signals that nobody could query efficiently. Organizations sitting on gold mines of information with no way to see it.
The goal was to build systems that could fix that. Pipelines that handle the volume. Transformation layers that impose structure. Frontends that make the output legible to people who aren't engineers. Over time, that turned into a practice — a specific way of approaching data problems that works across industries and scales.
Amicus Data is a one-person operation, and that's intentional. Every pipeline, every dashboard, every line of code comes from the same person who understood the problem. No handoff. No interpretation layer. No game of telephone between "the people who understand the data" and "the people who build the thing."
The focus is professional services — legal, accounting, medical — because those are industries where the data is rich, the competition is real, and the existing tools consistently fall short of what's actually needed.
What this looks like in practice.
The primary platform currently running ingests 660 million records per month from dozens of data sources, processes them through custom transformation pipelines, and surfaces the results in interactive analytics dashboards built for specific industries. Here's what that involves:
Collection at Scale
Data arrives from search APIs, mapping platforms, website crawlers, business directories, and public records. The volume is large — hundreds of millions of records — but the pipelines are built for it. They run on schedule, handle failures gracefully, and scale without architectural changes.
Transformation & Intelligence
Raw data gets cleaned, normalized, deduplicated, and enriched. AI classification models analyze content quality, detect competitive shifts, and score relevance. What goes into the warehouse isn't raw data — it's structured intelligence ready to be queried and visualized.
Applications & Dashboards
The final layer is the one people actually see: interactive web applications, data visualizations, competitive analysis tools, and market monitoring dashboards. Each one designed around the specific questions its audience needs answered.