Services
Four capabilities, scoped against the problem — not sold from a rate card.
EthanCorp works as an independent consultancy: project engagements or ongoing contracts, engaged directly by the people who own the problem. Every engagement states its constraints and its likely failure points before the build starts, not during it.
Services at a glance
Enterprise data integration & migration
Source systems that agree with each other, on a schedule you can state out loud.
BI & executive reporting
One number per metric, refreshed on a cadence that matches the decision it supports.
Data architecture & analytics engineering
A platform where a new report is a day of modelling, not a new integration project.
AI workflow automation
Hours of routine handling collapsed into minutes, with the judgement step kept and auditable.
Executive reporting and KPI standardisation are part of BI & executive reporting rather than a separate line item — in practice the dashboard and the definition are the same piece of work, and splitting them is how organisations end up with a governed model nobody uses.
Service 01
Enterprise data integration & migration
Connect ERP, MIS, CRM and cloud platforms into governed, observable data flows — and move data between them without a year of manual reconciliation.
The problem it solves
Every system holds part of the truth and none of them agree. Data moves by export, email and hand-edited spreadsheet, so nobody can say when a number was last correct — or whether last night's load actually finished.
A typical engagement
Typically 6–14 weeks for a first integration domain: discovery and source profiling, contract and schema design, pipeline build, reconciliation harness, then handover with runbooks. Migration work is scoped separately against the cutover date.
What gets delivered
- Source-to-target mapping and data contracts, written down and version-controlled
- Pipelines with retries, idempotent loads, and failure alerting that names the failing step
- A reconciliation harness that proves the target matches the source, run after every load
- Runbooks and a handover so the pipelines survive without me
Expected outcome
Source systems that agree with each other, on a schedule you can state out loud.
Technologies typically involved
- Talend
- SSIS
- Azure Data Factory
- Apache Airflow
- Python
- SQL
- REST/OpenAPI
- n8n
Service 02
BI & executive reporting
Replace the monthly spreadsheet assembly line with dashboards leaders actually open — on definitions the business has agreed to, not definitions each report invented.
The problem it solves
Three reports, three revenue figures, and a leadership meeting that spends its first twenty minutes arguing about which is right. Meanwhile someone rebuilds the same workbook by hand every month.
A typical engagement
Usually 4–10 weeks: metric inventory and definition workshops, semantic model, dashboard build against real decisions, then an adoption phase where the old spreadsheet is actually retired. KPI standardisation is part of this work, not a separate purchase.
What gets delivered
- A KPI dictionary — each metric with one owner, one definition, one calculation
- A governed semantic model so every report reads the same measures
- Executive and operational dashboards built around named decisions, not available fields
- A documented refresh cadence, and monitoring for when it breaks
Expected outcome
One number per metric, refreshed on a cadence that matches the decision it supports.
Technologies typically involved
- Power BI
- Tableau
- Looker Studio
- SAP BusinessObjects
- dbt
- SQL Server
- SAP HANA
Service 03
Data architecture & analytics engineering
Design the models, pipelines and semantic layers underneath the reporting — so the next ten questions do not each need their own bespoke extract.
The problem it solves
Reporting was built one urgent request at a time. Now there are forty extracts, no lineage, and every schema change breaks something nobody can find until month-end.
A typical engagement
4–12 weeks depending on estate size: current-state architecture review, target model design, incremental migration of the highest-traffic domains, and tests that run in CI. I prefer to prove the pattern on one domain before rolling it wider.
What gets delivered
- A documented target architecture with the trade-offs stated, not just the diagram
- Dimensional or semantic models built for the questions actually being asked
- Tested, version-controlled transformations with lineage you can follow
- A migration path that keeps the old reporting running until the new one is trusted
Expected outcome
A platform where a new report is a day of modelling, not a new integration project.
Technologies typically involved
- dbt
- SQL Server
- SAP HANA
- SAP BW
- Microsoft Fabric
- Azure
- AWS
- GCP
- SingleStore
- Git
Service 04
AI workflow automation
Automate the repetitive analytical work — document handling, report drafting, classification, retrieval — with APIs, Python, LLMs and orchestration, and a human review step where the decision matters.
The problem it solves
Skilled people spend their week re-keying, re-formatting and re-checking the same documents. AI is obviously relevant, but nobody wants a black box making decisions the organisation has to stand behind.
A typical engagement
Usually 6–12 weeks: pick one workflow with a measurable clock on it, instrument the current process, build the automation with an explicit human-in-the-loop gate, then measure the before and after. I will tell you the token and rate-limit cost before we start, not after.
What gets delivered
- One workflow automated end to end, with the manual baseline measured first
- Retrieval grounded in your own curated material — not the model's general memory
- An explicit review step, with accepted/rejected outcomes logged
- Usage and cost visibility, so the running bill is never a surprise
Expected outcome
Hours of routine handling collapsed into minutes, with the judgement step kept and auditable.
Technologies typically involved
- Python
- OpenAI
- Claude
- Gemini
- FastAPI
- Qdrant
- PostgreSQL
- n8n
- Docker
Engagement model
Project-based, or an ongoing contract.
Both start the same way. Scope, price and expected constraints are confirmed in writing after a paid discovery — never estimated from a first call.
Project engagement
A defined outcome with a start and an end: an integration domain, a reporting modernisation, one automated workflow. Fixed scope, written deliverables, handover with runbooks.
Best when you know what is broken and want it fixed and handed over.
Ongoing contract
A recurring allocation for a data function that needs senior capacity rather than a project: architecture decisions, pipeline ownership, roadmap, and support for an internal team.
Best when the work is continuous and you need it to stay maintainable.
Technical depth
What the delivered systems run on.
Listed because buyers ask, and because an estate that already uses these is an estate I can be productive in quickly. The tool is never the reason to do the project.
Languages & data engineering
- SQL
- Python
- Java
- dbt
Integration & orchestration
- Talend
- SSIS
- Apache Airflow
- Azure Data Factory
- n8n
- REST / OpenAPI
Databases & warehouses
- SQL Server
- Oracle
- MySQL
- SAP HANA
- SAP BW
- SingleStore
- PostgreSQL
Cloud & data platforms
- Microsoft Fabric
- Azure
- AWS
- Google Cloud
BI & analytics
- Power BI
- Tableau
- SAP BusinessObjects
- Looker Studio
AI & automation
- OpenAI
- Claude
- Gemini
- RAG / Qdrant
- Python automation
- Human-in-the-loop review
Enterprise systems
- SAP ERP
- Dynamics 365 / CRM
- HubSpot
- iSAMS
- PowerSchool
DevOps & delivery
- Git
- GitHub Actions
- Azure DevOps
- Docker
- CI/CD
Before you ask
Common questions.
Does EthanCorp work with a team or as an individual?
EthanCorp is a one-person consultancy operated by Dat Tran. The person who scopes the work is the person who builds it and hands it over. Where an engagement genuinely needs more hands, that is said before the contract, not discovered during it.
What does a typical engagement cost and how long does it run?
Engagements are scoped per project rather than sold from a rate card. A first integration domain is typically 6 to 14 weeks; a BI modernisation 4 to 10 weeks; a single automated workflow 6 to 12 weeks. Scope, price and the constraints expected are confirmed in writing after a paid discovery.
Can EthanCorp work alongside an existing internal data team?
Yes — that is the common case. The work is designed to be handed over, with runbooks and version-controlled models, so an internal team can own it afterwards rather than depending on an external contractor indefinitely.
Is AI required to work with EthanCorp?
No. Most of the delivered value is integration, data modelling and reporting governance. AI is applied where a repetitive workflow has a measurable clock on it and a human review step is acceptable — not as a default.
Not sure which of these you need?
Describe the symptom rather than the solution — a reporting cycle that is too slow, two systems that disagree, a team buried in manual handling. Scoping it correctly is part of the job.
- Response time
- Within two business days
- dattran.bi@gmail.com
- Based in
- Ho Chi Minh City, Vietnam — working across Asia and remote