FOLIO № 0002 / CHANNEL · ROBOTICS SOFTWARE

Khuzair Askyer.

Mechatronics Engineer / Robotics Software Developer

I build optimised robotic software focused on deterministic behaviour, real-time performance, and clean system integration — the kind of code that has to keep its promise to a moving machine.

OPEN TO WORK CV.PDF ↗ GITHUB ↗ LINKEDIN ↗ MAIL ↗

VITALS

Baseİstanbul

FieldAutonomous Robotics

StackROS 2 · C++ · Python

Exp.3+ yrs · robotics

EduSakarya UAS · Mechatronics · Y3

StylePerformance-first

Loop● closed

SIGNATURE

sha256 · askyer.k · 2c41 · 9f7a · 18bd · b0c2 · 5dd1 · ee93 · 7042

ROS 2 · jazzy · daily driver TARGET · NVIDIA Jetson Orin NX RT-LINUX · PREEMPT_RT · jitter < 80 µs CURRENT · GPU-accelerated perception OPEN FOR · contract · remote · İstanbul WRITING · in C++20 · in Python 3.12 ROS 2 · jazzy · daily driver TARGET · NVIDIA Jetson Orin NX

§ 01 ——

Specialisation.

Where I do my best work

DOMAIN

DESCRIPTION

STACK

LEVEL

S·01

Autonomous robotic systemsFull pipeline · perception → action

End-to-end autonomy stacks — perception, planning, control — composed into a single deterministic loop. Production-grade ROS architectures, scalable infrastructure.

ROS 2 · Nav2
C++17/20
state machines

core

▸

S·02

Real-time control softwareHard deadlines · low jitter

Deterministic control loops on PREEMPT_RT Linux — lock-free queues, priority-tuned threads, bounded allocations. Loops that hold their period under load.

RT-Linux · C++
lock-free
cyclictest

core

▸

S·03

SLAM & sensor fusionLocalisation under uncertainty

Visual / LiDAR / IMU fusion for autonomous platforms. Tight time-sync, calibration pipelines, drift-bounded estimation, and pose graphs that survive bad lighting.

OpenCV · PCL
EKF · pose-graph
OpenGL viz

core

▸

S·04

Embedded roboticsARM · Jetson · edge AI

From ARM MCUs to Jetson-class SoCs. GPU-accelerated perception on the edge, board bring-up, driver glue, memory-conscious processing under thermal and power limits.

Jetson · ARM
CUDA · TensorRT
edge-AI

core

▸

S·05

Low-latency system architectureMulti-threaded · async pipelines

Modular, fault-tolerant robotics pipelines — async processing, back-pressure, watchdogs, graceful degradation. Production-oriented engineering, clean integration.

async · DDS
Node.js bridges
monitoring

core

▸

§ 02 ——

Core stack.

Read-outs · self-reported

Languages & libraries.

C++ 17 / 20primary

ROS / ROS 2primary

Python 3.8 +daily

C — embeddeddaily

OpenCVin production

OpenGL — visualisationin production

JavaScript / Node.jsbridges · tools

CUDA / TensorRTedge perception

Hardware & systems.

EDGE COMPUTE

JetsonOrin · Xavier · Nano

GPU-accelerated perception · edge AI

EMBEDDED

ARMCortex-M / A

bring-up · drivers · BSP

RT-LinuxPREEMPT_RT

deterministic · low-jitter

MIDDLEWARE

ROS 2DDS · jazzy

production architectures

PERCEPTION

VisionOpenCV · CUDA

multi-cam · LiDAR · IMU

CONCURRENCY

Asyncmulti-thread

lock-free · fault-tolerant

§ 03 ——

Engineering profile.

What I optimise for

LATENCYMEMORYCONCURRENCYARCHITECTUREINTEGRATIONRELIABILITY

Deterministic behaviour

Memory efficiency

Asynchronous processing

Production-grade ROS

Scalable infrastructure

Fault-tolerant pipelines

§ 04 ——

Instrumentation.

Tools, by channel

Q-01 / STACK

Languages I think in.

C++ 17 / 20
Python 3.8 +
C — embedded
JavaScript / Node.js
ROS / ROS 2 — daily

Q-02 / SPECIALISATION

Where the code does work.

Autonomous robotic systems
Real-time control software
SLAM & sensor fusion
Embedded robotics
Multi-threaded pipelines

Q-03 / HARDWARE

Targets & environments.

NVIDIA Jetson
ARM-based embedded systems
Edge-AI robotics platforms
Real-time Linux (PREEMPT_RT)
OpenCV · OpenGL · CUDA

Q-04 / STYLE

How I write software.

Performance-first
Real-time aware
Hardware-conscious
Modular architecture
Production-oriented

§ 05 ——

A run, this morning.

ros 2 launch · autonomy bringup

askyer@jetson-orin — ros2 launch09:14 · +03

askyer@orin:~/ws_autonomy$ ros2 launch autonomy_bringup full_stack.launch.py rt:=true [09:14:02] ● rt scheduler · SCHED_FIFO 95 · cpu-isolated 2,3 [09:14:02] ● dds · cyclonedds · loopback OK · best-effort QoS [09:14:03] ● sensors · stereo@30Hz · livox@10Hz · imu@400Hz · gnss@10Hz [09:14:04] ● perception · gpu accel · tensorrt fp16 · 12.4 ms / frame [09:14:04] ● slam (lio-sam) · loop closure armed · cov ↓ [09:14:05] ● planner · nav2 · costmap 0.10 m · controller MPPI 50 Hz [09:14:05] ● watchdogs · 7 active · grace 200 ms [09:14:07] · loop period jitter peak 74 µs mean 11 µs (n = 100k) [09:14:31] ! dropped frame on cam_left (1) — recovered · log → /var/autonomy [09:18:48] ● waypoint 14 / 14 · path length 412.6 m · cte ≤ 0.08 m [09:18:49] → commit 7c1a·f0 — "bound allocations in cost layer" # note — the only metric I trust is the loop holding its period # under every condition the robot will actually see. # everything else is decoration.

§ 06 ——

Current interests.

What I’m reading / building toward

№

TOPIC

NOTE

I·01

Autonomous navigation.Robotics · planning

Long-horizon planning under partial information — costmaps, predictive control, and how a robot should behave when it isn’t sure.

ACTIVE ▸

I·02

AI-integrated robotics.Perception + control

Tightly integrating learned models inside deterministic control loops — what to learn, what to keep classical, where the seams should live.

ACTIVE ▸

I·03

GPU-accelerated perception.CUDA · TensorRT · Jetson

Squeezing real-time perception pipelines onto edge SoCs — bandwidth, kernels, fused pre-processing, and the discipline of zero-copy.

ACTIVE ▸

I·04

Embedded optimisation.ARM · RT-Linux

Profiling, cache awareness, bounded allocations, lock-free queues. The small disciplines that turn “it usually works” into “it always holds its period”.

ACTIVE ▸

I·05

Real-time robotics middleware.ROS 2 · DDS

Modern robotics stacks — QoS profiles, composable nodes, intra-process pipelines, and what production middleware should actually guarantee.

ACTIVE ▸

§ 07 ——

Correspondence.

RX · 48h · lab hours permitting

If you are building a robot that needs real-time software —
let’s talk.

CONTRACT FULL-TIME REMOTE / İSTANBUL CONSULTING

Mailhuzarh44@gmail.com
Phone+90 534 718 94 25
Baseİstanbul · TR
FieldAutonomous robotics
Experience3+ years
EducationSakarya UAS · Mechatronics · Y3
GitHub@huzarh ↗
LinkedIn/in/uzeyiraskyer ↗
StackROS 2 · C++ · Python
TargetJetson · ARM · RT-Linux
StylePerformance-first