services

Equipping teams without taking their decisions for them.

Six services with one requirement: everything an engagement produces must be readable, reproducible and approvable by an in-house engineer. Without a handover procedure the client inherits files it cannot use. A generated deliverable nobody can explain causes the same problem.

01services
typical duration · 6 months to 1 year
build
web apps · analytics · machine learning
Web applications, analytics platforms and machine learning services, from prototype to production on AWS or Azure.
typical duration · 3 to 6 months
migrate and modernise
platform migration · legacy code
Cloud platform migration and takeover of existing code. Services that must keep running during the transition stay in service.
typical duration · 3 months
audit and govern
system audit · data governance
Critical review of a system already in place, and data governance the team keeps applying after the engagement.
typical duration · 6 months to 1 year
hand over and train
new platform · adoption · capability building
Adoption of a new platform by the team that will run it: a use case built with them, the architecture explained, training on their own data. The engagement ends when the in-house team delivers the next dashboard without help.
typical duration · 6 months to 1 year
adopt AI and change how teams work
generative ai · agents · use cases
Adoption of generative AI and agents: use cases chosen on measured impact, written rules of use, team training. The quality of the result depends mostly on the data given to the model, and what changes is how the business teams work.
create an online presence
association · liberal profession · independent
Visual identity and online presence for a small organisation. The deliverable includes the design system (tokens, components, written rules) as well as the pages, so the organisation can develop its site without a supplier. This site is one example.
A hand-drawn land survey: columns of accounts at the left, a plan of lots, roads and houses at the right, fold marks and stains of use.
Anonyme, Nouvelle-AngleterreSurvey of Sturgeon Creek, Maine, 4 July 1752
02technical stack

The technologies used in production and the role of each. The choice depends on the client’s constraints. A stack the in-house team cannot maintain ends up being replaced, at its own cost.

compute and data

Storage, query, transformation and reporting. The choice depends on the client’s constraints: an existing cloud contract, sovereignty requirements, in-house skills.

aws · query over files
Athena

A SQL query engine that reads files stored on S3 directly, with no database to load and no server to maintain. Billing depends on the volume of data scanned, which makes partitioning a design decision to take from the start.

aws · distributed compute
EMR

Hadoop, Spark and Trino clusters created on demand, for workloads too large for a single machine. Existing Spark code runs there without a rewrite, which is often the reason for choosing it.

azure · event-driven execution
Functions App

Azure’s serverless service: code triggered by an uploaded file, a schedule or an HTTP call, with no server to size. Suited to short processing steps; for long-running jobs, per-second billing makes it expensive.

microsoft · relational database
SQL Server

Microsoft’s relational engine, widespread in enterprise systems. In a migration it is almost always the source to read and the constraint to respect, rarely the target.

open source · relational database
PostgreSQL

The default target when nothing requires a proprietary engine: open licence, portable across clouds. Its extensions, PostGIS for geospatial data and pgvector for vector search, avoid adding a second system for a single need.

language models · api
OpenAI

The GPT models, called through OpenAI’s API or through Azure OpenAI when the client already has an Azure contract. Used for extraction, classification and writing assistance. Engineering decisions stay with an engineer.

anthropic · language models
Claude

Anthropic’s Claude models, called through Anthropic’s API or through AWS Bedrock, Google Vertex AI and Microsoft Foundry. The client’s cloud provider therefore does not dictate the choice of model. Used for document analysis, writing assistance and agents that call tools.

microsoft · github · assistants
Copilot

Two products under the same name: GitHub Copilot for code assistance in the editor, and Microsoft Copilot with Copilot Studio for business agents on Azure. The agents built at Egis Group are based on the latter.

data warehouse
Snowflake

A data warehouse that separates storage from compute: each is sized separately, and compute is billed per use. Available on all three major clouds, which avoids tying the warehouse to the infrastructure provider.

aws · managed models
Bedrock

Several model families available through a single API, without data leaving the AWS account. The choice of model is recorded as an architecture decision, open to revision.

microsoft · integrated platform
Fabric

Ingestion, warehouse, notebooks and reporting in one product, on shared storage (OneLake). The integration saves time; in return, the platform ties the client to the Microsoft ecosystem. That trade-off is settled at scoping.

microsoft · reporting
Power BI

Semantic models, reports and dashboards, used directly by business teams. Ambiguous metric definitions become visible there, which makes it a good test of the quality of the data model upstream.

languages

Three languages, chosen because an in-house team can read them with no special training.

processing · models
Python

The language of data processing and machine learning, from the ingestion script to the model service. pandas, PySpark and scikit-learn cover every step without changing tools.

query · transformation
SQL

The only language read by both business and technical teams. A transformation written in SQL stays readable years later, which is rarely the case for a pipeline written in code.

declarative infrastructure
Terraform

Describes cloud infrastructure in versioned files, applied reproducibly. Without it, a platform’s configuration exists only in a web console, where it cannot be reviewed, handed over or audited.

cloud providers

Two providers used in production. Google Cloud is covered by certifications but not listed here, for lack of a recent engagement to cite.

amazon web services
AWS

The widest catalogue of data services: S3, Athena, EMR, Bedrock. It is the default choice when the company does not already have a Microsoft contract.

microsoft
Azure

The natural provider for companies already on Microsoft 365 and Entra ID, where identity, billing and compliance are already in place. Fabric, Functions App and Azure OpenAI are its most common data services.

03method

Every engagement follows the same four steps. Decisions are taken at scoping, and that is the step that is tracked. The scope is written down so that nobody waits for an access, an environment or a decision.

survey
Interviews, sources, real constraints
scoping
Written scope, what is excluded and why
construction
Incremental deliverables, measured
handover
In-house team, sole signatory
04terms
first stepa 5 to 10 day scoping engagement, after which the next step is decided or not
formatfixed-price or time-and-materials, 3 to 5 days a week
duration3 months, 6 months or 1 year depending on the service
rate€550 to €750 per day excluding VAT, depending on the project
locationParis and Île-de-France · remote elsewhere
invoicingSASU Commutator · VAT FR73979139037
what is ruled out

Requests declined, whatever the length of the engagement.

  • An engineering decision automated with no accountability framework.
  • A deliverable no in-house engineer can read.
  • A figure announced without a measurable source.
  • Scaling up before a pilot has run in production for three months.