Numerical Compute Kernels
Custom C/C++ kernels for linear algebra, FFTs, ODE solvers, optimisation, and image/signal processing — exposed to MATLAB via MEX, with SIMD/BLAS acceleration where it matters.
MATLAB & Simulink Core Engine Development
Custom MATLAB Toolboxes, Simulink Models & Embedded Codegen
We engineer custom numerical compute kernels, MATLAB toolboxes, Simulink S-functions, MEX bridges, and embedded C/C++ replacements for control systems, signal processing, model-based design, and simulation workloads — for automotive, aerospace, energy, robotics, and research teams that need production-grade speed and traceability.
MEX · Simulink Coder · MISRA · AUTOSAR · HIL · Senior engineers
C/C++MEX · S-Function
SimReal-time HIL
AUTOSAREmbedded targets
100%Spec-traceable
MATLAB & Simulink Solutions We Build
From custom numerical kernels to production-ready embedded codegen — engineered with the rigour automotive and aerospace processes demand.
MATLAB Toolboxes
Domain-specific MATLAB toolboxes — control systems, communications, biomedical, finance — with documented APIs, examples, and unit tests engineering teams trust.
Simulink S-Functions & Blocks
Custom Simulink blocks via S-functions (level-2 MATLAB and C MEX) — for hardware models, plant models, and proprietary algorithms with proper sample-time semantics.
Embedded Code Generation
Production-ready embedded C/C++ generated via Simulink Coder / Embedded Coder — AUTOSAR-classic and adaptive targets, MISRA-aware, with custom storage classes and timing budgets.
HIL / SIL Test Rigs
Hardware-in-the-Loop and Software-in-the-Loop rigs using Speedgoat, dSPACE, or custom RT targets — plus test harnesses, fault injection, and signal logging.
MATLAB-to-Native Migration
Re-implement MATLAB prototypes as production C++/Rust/Python services when MATLAB runtime, licensing, or deployment topology no longer fits.
Industries We Serve
Engineering teams in automotive, aerospace, energy, MedTech, robotics, and quant finance — wherever models meet production hardware.
Automotive
Aerospace
Energy & Power
MedTech
EdTech / Research
Quant Finance
Robotics
DefenceTech
Stack & Tooling
Modern MATLAB/Simulink platforms, codegen toolchains, HIL targets, and static-analysis tools that engineering organisations actually run.
MATLAB R2024+Platform
SimulinkPlatform
Simulink CoderCodegen
Embedded CoderCodegen
C / C++ (MEX)Bridge
StateflowState Machines
AUTOSARAutomotive
TargetLinkCodegen
Speedgoat / dSPACEHIL
Polyspace / MISRAStatic Analysis
Python / NumPyMigration
Eigen / BLASLinear Algebra
Our Model-Based Engineering Process
A six-step rhythm engineered so models survive Tier-1 audits, codegen lands on target hardware, and HIL closes the loop.
Model & Algorithm Audit
We review your existing models, requirements (often in DOORS or Polarion), sample-time architecture, fixed/floating point posture, and target hardware constraints.
Architecture & Sample Times
Decomposition into model references and libraries, signal naming conventions, sample-time strategy, atomic subsystems, and traceability links to requirements.
Implement & Verify
Model implementation, S-functions and MEX bridges as needed, unit tests via Simulink Test, requirement-based test cases, and coverage measurement.
Codegen & MISRA
Embedded Coder configuration, MISRA-C compliance via Polyspace or custom rule packs, custom storage classes, and timing/memory budget verification.
HIL / SIL Validation
Software-in-the-Loop and Hardware-in-the-Loop testing on Speedgoat or dSPACE rigs, fault injection, and back-to-back comparison against the model.
Handover & Support
Documentation, training, version-control workflows (Git + Simulink Project), and a retainer for releases, MATLAB version upgrades, and new feature work.
Why Choose Hendoi for MATLAB & Simulink
Six commitments that decide whether your model-based engineering project ships safely — or stalls in code-review limbo.
Real Engineering Discipline
Requirement traceability, MISRA compliance, sample-time correctness, and unit-tested models — the disciplines that decide whether your software passes a Tier-1 OEM audit.
C/C++ + MATLAB Bilingual
Our engineers move comfortably between MATLAB/Simulink and C/C++. We can write a MEX, hand-write an embedded driver, and review the generated code — all in the same week.
Migration Honesty
We will tell you when MATLAB is the right home and when production really should be C++/Rust/Python — including a transparent migration plan and risk register.
AUTOSAR & Safety-Aware
Familiarity with AUTOSAR Classic and Adaptive, ISO 26262 process expectations, and safety-critical workflows — without claiming certifications we do not hold.
Senior Engineers Only
Model-based engineering rewards experience. Our seniors have shipped models that ended up in production vehicles, drones, and medical devices.
Transparent Engagement
Weekly demos with model artefacts and HIL runs, direct access to engineers, signed SOWs with sample-time and target assumptions documented.
Engagement Models
Pick the shape that matches whether you need a focused algorithm sprint, a programme squad, or migration off MATLAB.
Algorithm / Kernel Sprint
Implement a specific algorithm or numerical kernel as a tested MATLAB function or MEX module, with documentation and example scripts — typically 4-6 weeks.
- Focused algorithm scope
- MEX or pure MATLAB
- Docs + tests + examples
Model-Based Engineering Squad
A senior squad — MBD, embedded C, test — building Simulink models, generating production code, and validating on HIL rigs as an extension of your engineering org.
- Senior MBD + embedded engineers
- Codegen + MISRA + HIL
- Requirement traceability
Migration & Replacement
Re-implement MATLAB code as production C++/Rust/Python services with numerical equivalence testing, performance benchmarks, and rollout support.
- Numerical equivalence tests
- Performance benchmarks
- Rollout + dual-run window
Real-World Use Cases
Representative MATLAB and Simulink projects engineered for EV, aerospace, industrial DSP, MedTech, energy, and quant finance.
EV Battery Management Model
Simulink BMS model with state-of-charge estimation, cell-balancing, and thermal management — Embedded Coder output deployed to an Infineon Aurix target with MISRA compliance.
Drone Flight Control
Cascaded PID + EKF flight controller for a research drone — modelled in Simulink, validated on HIL, and code-generated to an STM32H7-based flight computer.
Industrial DSP Toolbox
Custom MATLAB toolbox for a vibration-monitoring OEM — adaptive filtering, order tracking, and condition indicators with documented APIs and example scripts.
Medical Imaging Kernel
C++ MEX module for fast 3D image registration in a diagnostic-imaging product — SIMD-accelerated and benchmarked against the previous MATLAB-only path.
Power-Grid Simulation Engine
Real-time grid simulation on a Speedgoat target for a transmission utility — protection-relay HIL testing, fault injection, and result-logging pipelines.
Quant Strategy Migration
Production migration of a MATLAB quant strategy to a Python + NumPy/Pandas stack with numerical equivalence tests and a 30-day shadow-run window before cutover.
Frequently Asked Questions
Common questions engineering managers ask before bringing an external team into model-based engineering work.
What MATLAB and Simulink versions do you work with?
MATLAB R2022a and newer (R2024a/b/R2025a most common). For older projects we can stand up the version matching your delivery line — version pinning is treated as a first-class requirement.
Do you generate production embedded code from Simulink?
Yes — Embedded Coder for automotive and industrial targets, including AUTOSAR Classic and Adaptive, with MISRA-C compliance verified via Polyspace and custom rule packs. We do not claim ISO 26262 certification — but we work cleanly inside customer ASIL processes.
Can you write MEX functions in C/C++?
Yes — level-2 MATLAB S-functions, C MEX functions, and C++ MEX with the new C++ MEX API. We pick based on performance needs, memory ownership semantics, and the target Simulink solver.
Do you do HIL testing?
Yes — Speedgoat, dSPACE, and bespoke real-time targets. We build the test harnesses, fault-injection scripts, signal-logging pipelines, and back-to-back comparison reports.
When should I migrate from MATLAB to a production stack?
When MATLAB runtime licensing, deployment topology, or runtime performance no longer fits. We migrate to C++/Rust/Python with numerical equivalence testing and a transparent dual-run window before cutover. Sometimes the right answer is to keep MATLAB — we will say so.
Do you integrate with our requirements management tools?
Yes — Requirements Toolbox links to DOORS, Polarion, Jama, and Codebeamer are common in our work. Traceability matrices are part of the delivery, not an afterthought.
How is licensing handled?
You retain your MATLAB and toolbox licences. We bring our own development licences for build/test work and never check generated code into your repo with licence-protected blobs. Customer licence keys never leave your environment.
How long does a MATLAB/Simulink project take?
Focused algorithm work: 4-8 weeks. Full model-based engineering programmes with HIL validation: 4-9 months depending on scope, target safety class, and customer process gates.