Senithu Dampegama — Computer Vision Engineer, Defence R&D
00 / INDEX SYSTEM / ACTIVE

SENITHU
DAMPEGAMA

COMPUTER VISION ENGINEER
DEFENCE R&D

I engineer real-time perception systems for autonomous and embedded platforms.

From datasets and neural networks to edge deployment and physical control.

CURRENT
SENTARI LABS
FOUNDER
AETRACH INDUSTRIES
BASED
UNITED KINGDOM
Senithu Dampegama standing in front of a bronze sculpture group
SUBJECT / SENITHU DAMPEGAMA ROLE / COMPUTER VISION ENGINEER DOMAIN / DEFENCE R&D SYSTEM / ACTIVE
FIG. 00 — SUBJECT {{ heroHint }}
OBJECT DETECTION/ TRACKING/ EDGE AI/ TENSORRT/ JETSON/ AUTONOMY
01 / PROFESSIONAL PRACTICE COMPUTER VISION — DEFENCE R&D

REAL-TIME PERCEPTION.
BUILT FOR DEPLOYMENT.

Developing and deploying real-time computer-vision systems for embedded platforms, from dataset engineering and model development through tracking, optimisation and system integration.

CURRENT ROLE SENTARI LABS / DEFENCE R&D / 2025 — PRESENT
PERCEPTION PIPELINE — SENSOR TO DEPLOYMENT {{ stageReadout }}

{{ stageBlurb }}

{{ ss }}
{{ c }}
02 / CAPABILITY MAP SIX-STAGE SYSTEM

FROM SENSOR
TO DECISION AT THE EDGE.

{{ stackName }}

{{ stackState }}

{{ stackBlurb }}

{{ t }}
DEMONSTRATED IN
03 / SELECTED ENGINEERING SYSTEMS ALL SYSTEMS →

PERCEPTION THAT
LEAVES THE LAB.

Human-following Ackermann robot v2 — completed platform with OAK-D Pro W stereo camera
TRACK LOCK DEPTH SOURCE / STEREO SENSOR / OAK-D PRO W
EDGE MAP / COMPUTED FROM FRAME
Ackermann steering gear train driven by a DC motor
STEERING GEAR TRAIN — MECHANICAL BIAS COMPENSATED IN CONTROL
Jetson Orin Nano and motor drivers mounted on the robot chassis
JETSON + DRIVE ELECTRONICS — COMPONENT PLACEMENT
FLAGSHIP 01 — SOLO SYSTEM / COMPUTER VISION + EMBEDDED CONTROL

HUMAN-FOLLOWING
ACKERMANN PLATFORM

A vision-guided follower robot rebuilt as a production-quality stack: on-camera spatial detection with stereo depth, stable multi-person tracking with target lock, and a safety-first decision layer that never moves without a confirmed target.

RGB + STEREO DEPTH ON-CAMERA SPATIAL DETECTION MULTI-PERSON TRACKING TARGET LOCKING FAULT RECOVERY
SIGNAL PATH
{{ n.name }}→
L5 DECISION STATE MACHINE SEARCH IS STATIONARY BY DESIGN
INIT SEARCH ACQUIRE FOLLOW FAULT → RECOVER
OPEN CASE STUDY
GAIT / CRAWL SEQUENCE / FL→RR→FR→RL
LOCOMOTION TEST — PHYSICAL PLATFORM
CRAWL GAIT — FL → RR → FR → RL
FL
FR
RL
RR
Quadruped locomotion assembly simulated in NVIDIA Isaac Sim
SIMULATED FIRST — NVIDIA ISAAC SIM
FLAGSHIP 02 — 5-PERSON TEAM / INDIVIDUAL OWNERSHIP SHOWN

QUADRUPED
BEHAVIOUR ARCHITECTURE

TEAM SIZE / 05
ROLE / SCRUM MASTER — SOFTWARE SPRINT
OWNERSHIP / CONTROL ARCHITECTURE · GAIT · LOCOMOTION CAD · REPOSITORY INTEGRATION

A five-person CDIO quadruped. I led the programming and behaviour-control sprint: an eight-module MicroPython firmware stack (~1,700 lines) with priority-based behaviour arbitration — predator response interrupts obstacle avoidance, phototaxis and walking mid-stride — plus the leg mechanism CAD and a crawl gait designed in NVIDIA Isaac Sim before touching hardware.

BEHAVIOUR ARBITRATION — PRIORITY ORDER
P0PREDATOR RESPONSE — HALT + DEPLOY WINGS
P1OBSTACLE AVOIDANCE
P2PHOTOTAXIS — STOP ~20 CM FROM SOURCE
P3LOCOMOTION — INTERRUPTIBLE WALK
OPEN CASE STUDY
MODE / WALL FOLLOW ESCAPE / EVERY ATTEMPT
Maze robot front — two ultrasonic sensors mounted at fixed 25 degree angles
TWO ULTRASONIC PODS — FIXED 25° MOUNTS
Fusion 360 CAD model of the maze robot
CAD — COMPACT FOOTPRINT CONSTRAINT
FLAGSHIP 03 — SOLO COURSEWORK SYSTEM

AUTONOMOUS MAZE
NAVIGATION

The maze was physically rebuilt every week, so the solution had to generalise. With only two ultrasonic sensors at fixed 25° angles and a strict footprint limit, a carefully tuned wall-following algorithm escaped the maze on every attempt.

SENSORS2× ULTRASONIC — NO CAMERA, NO LIDAR
GEOMETRYFIXED 25° MOUNTS
FOOTPRINTSTRICT MAXIMUM ENVELOPE
RESULTESCAPED ON EVERY ATTEMPT

Robust autonomy does not always require a neural network — it requires the right algorithm for the sensing you actually have.

OPEN CASE STUDY
LABORATORY / EARLIER SYSTEMS
04 / FOUNDER

AETRACH
INDUSTRIES

Building an advanced engineering company around intelligent, autonomous and embedded systems.

FOUNDER / TECHNICAL DIRECTION / UNITED KINGDOM
ENTER AETRACH INDUSTRIES VIEW FOUNDER PROFILE

AETRACH is the organisation. This site is the engineer behind it — the technical evidence, project history and professional identity of its founder.

05 / EXPERIENCE OPERATIONAL RECORD
2025 — PRESENT

COMPUTER VISION ENGINEER — DEFENCE R&D

SENTARI LABS

Developing lightweight object-detection and tracking systems for embedded counter-UAS applications, with work spanning dataset engineering, model evaluation, optimisation and Jetson deployment.Some professional work cannot be documented publicly.

2026 — PRESENT

FOUNDER & ENGINEER

AETRACH INDUSTRIES LTD

Technical direction and engineering development across computer vision, autonomous systems, robotics and edge computing, with a long-term aerospace and defence focus.

2023 — 2026

ENGINEERING ASSISTANT

SURVEYING / LIDAR / 3D SCANNING

Leica terrestrial LiDAR systems and BLK ARC. Point-cloud capture in tunnels and infrastructure. AI-assisted scanning workflows.

ROBOTICS & AI TUTORTeaching robotics and AI fundamentals.
FREELANCE ENGINEERINGRobotics and AI engineering for private clients.
TECHNICAL / ITEarlier technical and IT roles.
06 / ABOUT ENGINEER PROFILE
Senithu Dampegama
FIG. 06 — FIELD CROP / 4:5

I work at the boundary between machine perception and physical systems.

My focus is taking computer-vision systems beyond experiments — into embedded hardware, autonomous platforms and real operating constraints.

COMPUTER VISION EDGE DEPLOYMENT AUTONOMOUS SYSTEMS EMBEDDED CONTROL ROBOTICS MECHANICAL CAD
EDUCATION GPA 4.44 / 4.5 — TOP 5% OF COHORT
BEng Robotics & Artificial Intelligence
UNIVERSITY OF HERTFORDSHIRE · 2024 — 2028
EMBEDDED SYSTEMS · MACHINE LEARNING · ROBOTICS CONTROL · AI ENGINEERING
ENGINEERING PRINCIPLES
{{ pr.num }} {{ pr.title }} {{ pr.body }} EVIDENCE →
07 / CONTACT CHANNELS OPEN

START A TECHNICAL
CONVERSATION.

DEFENCE R&D/ COMPUTER VISION/ AUTONOMOUS SYSTEMS/ EDGE AI/ ENGINEERING COLLABORATION
VIEW CV DOWNLOAD PDF PDF · 4 PAGES
{{ email }} ADDRESS COPIED / CHANNEL OPEN
DIRECT TRANSMISSION {{ formStateLabel }}
MESSAGE SENT SUCCESSFULLY — A COPY HAS BEEN SENT TO YOUR EMAIL ADDRESS. MESSAGE NOT SENT — THE CONNECTION FAILED. PLEASE RETRY. {{ formInvalidMsg }}