evonitec
DE

Focus · Semiconductor & test

Where our way of working came from.

Semiconductor and ATE test, radar sensing, wireless and test & measurement: 25 years in an environment where test time is optimised in milliseconds and yield is tracked to the decimal point. The sections below go into that level of detail.

Arrange an initial callSee the tools
Pick-and-place handler removing tested dies from a wafer

Situation

The technology is ready. Day-to-day adoption is not.

The major platform vendors have long been shipping what AI on the test floor requires: compute right at the tester, so data can be evaluated while the lot is still running; data platforms that make test results available without delay; and assistants that support test program development. The gap is not in those building blocks but between them — in the question of who brings them together in everyday engineering, operates them and owns the result.

Typical picture: hiring freeze while project pressure remains · a grown landscape of datalogs and reports with a zoo of spreadsheets and scripts · many isolated AI initiatives but no line across departments · customer IP and export control limit what may leave the building · site relocation forces knowledge transfer and standardisation.

Three building blocks

How an entry point realistically begins.

Block 01

Assessment and roadmap for the test floor

Inventory of the data landscape — datalogs, test results, connection to the manufacturing system. From that, three to five prioritised use cases with effort and benefit estimates plus a platform recommendation.

→ Clearly bounded start-up effort, costed up front
Block 02

AI-assisted test program development

Introducing assistant and agent workflows into program development — with review gates, traceability and team enablement. Without governance you get exactly the review bottleneck that AI was supposed to remove.

→ Pilot on a real test program
Block 03

Test time and yield optimisation

Adaptive test limits, outlier screening following documented methodology, data-driven test flow reduction, analysis close to the tester. Success is measurable.

→ Metrics: test time and ppm / yield
AI and data competence → Test depth and proximity to manufacturing → Classic test service providers Technical depth present — analysis missing Capacity instead of method Data science consultancies AI competence present — technical depth missing Have to learn the test floor first evonitec Test floor practice + AI + project leadership vendor-independent, available on short notice Platform vendors' own teams strong, but platform-bound and often not available
Most players in the market have exactly one of the three ingredients. The combination of test floor practice, AI competence and project leadership experience is rare — and tied to no platform vendor.
Vendor-independent — deliberately. Analysis close to the tester, adaptive test limits and predictive binning rest on the same concepts across all major platforms. Methodology and recommendations therefore hold platform-agnostically. Which environment you run changes the specific implementation — not the underlying method.

Further fields

Radar, wireless and test & measurement.

Radar sensing

Concept and system development, production test of highly integrated sensors, signal analysis and evaluation.

Wireless and IoT

Mobile platforms, connectivity, energy-efficient applications — from product definition and portfolio planning.

Test & measurement

Test systems, test specification, measurement data analysis and system architecture — from development to production operation.

Hardly any project starts on a green field. The typical task is integration into an environment that grew over time: an existing data landscape, evaluations with history, scripts with a past. Starting there, without triggering a platform migration, is a realistic first step — and usually the most economical one.

Where would you start?

A short description of your situation is usually enough to clarify whether and where a first step is worthwhile. No obligation — by email or by phone.

Write an email+49 8122 558351