Platform & Architecture

One platform. Three layers. One consistent stack.

talenttrends is built as a layered architecture — the Data Hub as the data foundation, the Integration Layer as the pipeline layer, the Micro Apps as the application layer. All three share the same technical stack, the same security controls and the same operations logic. What you’ll find here is the architecture that supports it.

talenttrends logo horizontal light gray

Architecture

Three lay­ers, clear responsibilities

Each of the three tal­ent­trends lay­ers has a clear­ly defined task. You can under­stand them indi­vid­u­al­ly, use them indi­vid­u­al­ly, oper­ate them indi­vid­u­al­ly — but they are built to work most pow­er­ful­ly on top of one another.

Data Hub

The data foun­da­tion. Cor­ner­stone data is repli­cat­ed into a rela­tion­al data­base of your choice — schema-sta­ble, queryable, exportable. The foun­da­tion for every­thing above it.

Integration Layer

The pipeline lay­er. On the basis of the repli­cat­ed data, map­ping, rout­ing and deliv­ery are defined — between Cor­ner­stone, Suc­cess­Fac­tors, Microsoft Graph and your own systems.

Micro Apps

The appli­ca­tion lay­er. Embed­ded in Cor­ner­stone, fed from the repli­cat­ed data foun­da­tion. Ready-made appli­ca­tions from the cat­a­logue or tai­lor-made via the pub­lic SDK.

Iden­ti­ty and authen­ti­ca­tion, role-based access, audit log­ging, mon­i­tor­ing and update process­es — all lay­ers share the same infrastructure.

Technology

Stan­dards that are at home in your IT

tal­ent­trends uses com­po­nents that are already present in most enter­prise IT land­scapes any­way. No exot­ic frame­works, no black-box libraries, no lock-in to data­bas­es, iden­ti­ty or inte­gra­tion inter­faces. What you get from this — and why it’s the right approach:

What you get
What it means
Microsoft enterprise stack

Estab­lished ecosys­tem, broad devel­op­er avail­abil­i­ty, long-term maintainability.

Choice of database

SQL Serv­er or Post­greSQL — local, in your cloud or with a sov­er­eign EU provider. No data­base lock-in.

Container distribution

Runs in any con­tain­er run­time, cloud-agnos­tic, hyperscaler-independent.

Standard identity

OAuth 2.0, SAML, OpenID Con­nect — inte­grat­ed into your exist­ing iden­ti­ty landscape.

Open integration protocols

ODa­ta v4, REST, SCIM, SFTP, ODBC — no pro­pri­etary interfaces.

EU-compliant AI

LLM infer­ence in EU regions, con­fig­urable sep­a­rate­ly from the platform.

Multi-Tenancy

One set­up, many ten­ants, no data mixing

tal­ent­trends is built from the ground up for mul­ti-ten­ant sce­nar­ios — whether you run a group-wide plat­form for sev­er­al sub­sidiaries or serve mul­ti­ple end cus­tomers as a reseller. Ten­ant bound­aries are enforced in every layer.

Clean tenant separation

Every record car­ries a ten­ant iden­ti­fi­er. Requests see only the data of the request­ed ten­ant — auto­mat­i­cal­ly, with­out appli­ca­tion log­ic hav­ing to enforce it itself.

A separate configuration per tenant

Cor­ner­stone por­tal, host­ing region, LLM mod­el, secu­ri­ty poli­cies, theme and lan­guage can dif­fer per ten­ant. A sin­gle plat­form instance can serve sev­er­al ten­ants in dif­fer­ent configurations.

Separate databases per tenant

The repli­cat­ed Cor­ner­stone data­base of each ten­ant runs in a sep­a­rate tar­get data­base — even in dif­fer­ent host­ing regions.

Scaling

From sin­gle ten­ant to sev­er­al pods, with­out an archi­tec­ture change

tal­ent­trends scales along two axes — ver­ti­cal (more resources per instance) and hor­i­zon­tal (sev­er­al pods in par­al­lel). Which vari­ant makes sense for you depends on your load pro­file, high-avail­abil­i­ty require­ments and host­ing model.

Vertical scaling

More CPU, more RAM, larg­er data­bas­es — the stan­dard route for most DACH enter­prise cus­tomers. Scales to data vol­umes of up to sev­er­al mil­lion rows per table with­out an archi­tec­ture change.

Horizontal scaling

Sev­er­al instances in par­al­lel, coor­di­nat­ed via a cen­tral con­fig­u­ra­tion lay­er. Rec­om­mend­ed for high load, geo­graph­ic dis­tri­b­u­tion or high avail­abil­i­ty across zones.

High availability

Mul­ti-node setups for the data­base (syn­chro­nous repli­ca­tion), appli­ca­tion repli­cas for the plat­form, auto­mat­ic recov­ery if an instance fails.

Disaster recovery

Auto­mat­ed back­ups with point-in-time restore. Recov­ery time objec­tives and recov­ery point objec­tives are defined per customer.

Extensibility

Not a closed plat­form. You build along.

tal­ent­trends is not built as a mono­lith­ic prod­uct, but as a plat­form with defined exten­sion points. Your IT, your con­sul­tants or third par­ties can devel­op their own com­po­nents — with­out talessio hav­ing to release code each time.

Public Micro Apps SDK

The tal­ent­trends Micro Apps SDK is avail­able as a pub­lic NuGet pack­age. It lets you devel­op your own Micro Apps against your repli­cat­ed data foun­da­tion — with the same secu­ri­ty, them­ing and life­cy­cle mech­a­nisms as the cat­a­logue apps.

Plugin architecture

New Micro Apps are hot-deployed as plu­g­ins, with­out a plat­form release. One plu­g­in file into the plu­g­in fold­er, auto­mat­ic detec­tion, avail­able immediately.

Open interfaces

REST APIs, ODa­ta end­points and web­hook con­ven­tions are doc­u­ment­ed. Any­one want­i­ng to run their own appli­ca­tions out­side the Micro App plat­form has all the hooks they need.

Operations

What runs can be observed

A plat­form is only as good as the oper­a­tions tools that make it oper­a­ble. tal­ent­trends comes with the tools you expect for secure oper­a­tion — whether you oper­ate it your­self or have us oper­ate it.

Structured logging

All logs in machine-read­able for­mat, with cor­re­la­tion IDs across lay­ers. Con­nectable to Splunk, ELK, Data­dog or your own log­ging system.

Health checks and metrics

Stan­dard­ised health end­points per ser­vice, Prometheus-com­pat­i­ble met­rics, clear indi­ca­tors for up/​down/​degraded.

Audit trail

Every con­fig­u­ra­tion-chang­ing action and every data access by end users is logged — with user, time­stamp, action, result. For inter­nal trace­abil­i­ty, for audits, for forensics.

Zero-downtime updates

Updates are rolled out via par­al­lel deploy­ment slots — the run­ning ver­sion stays active until the new ver­sion is verified.

Backup and restore

Auto­mat­ed back­ups on a sched­ule, point-in-time restore, doc­u­ment­ed recov­ery pro­ce­dures. Restore tests are part of our operations.

Monitoring of sync pipelines

Every sync run of the Data Hub and every Inte­gra­tion Lay­er pipeline is record­ed with start, end, vol­umes and errors. You see the state of every data flow with­out hav­ing to dig into the logs.

20230323 131718 087 talessio scaled

Interested?

Talk to our architects

90 min­utes, deep dive. We’ll go through the archi­tec­ture with your archi­tects — lay­er mod­el, stack, exten­sion points, oper­a­tions. With your spe­cif­ic ques­tions, your archi­tec­tur­al con­cerns, and the lev­el of depth you need.

Arrange a call

hello@​talessio.​com