RETURN_TO_BASE

ARCHIVE_VOL.3 / LAB

Labs

Active

A collection of experiments, prototypes, and technical explorations across AI, robotics, data, and software engineering.

Things I'm currently building, testing, and trying to understand.

Showing 2 of 2 experiments

Gesture-Controlled Drone — work in progress
EXP_01In ProgressAUG 2026
RoboticsComputer VisionAI / ML

Gesture-Controlled Drone

A prototype exploring gesture-based control of a programmable drone — investigating how computer vision and gesture recognition can translate human hand movements into real-time commands for a physical robotic system.

Objective

Translate hand gestures into real-time flight commands for a physical drone.

Current Approach

Computer-vision-based gesture recognition running against a programmable drone platform, with an emphasis on keeping recognition latency low enough for responsive, real-time control.

What I'm Testing

How large a gesture vocabulary the system can reliably distinguish, and where the recognition pipeline needs to run to keep latency within a usable range.

Current Milestone

Building the real-time gesture recognition pipeline and testing it against onboard flight controls.

Next Step

Expanding the gesture vocabulary and tightening recognition latency.

Concepts & Technologies

Computer VisionGesture RecognitionReal-Time ProcessingRoboticsEmbedded SystemsAI
MORE DETAILS ON EXPERIMENT SOON
hushHush Recruiter — work in progress
EXP_02In ProgressAUG 2026
AI / MLSystemsSoftware Engineering

hushHush Recruiter

An AI-driven recruitment automation project exploring automated candidate discovery, evaluation, and hiring workflows — analysing publicly available developer signals to help surface and evaluate candidates.

Objective

Explore whether public developer signals can meaningfully inform candidate discovery and ranking, and what an automated, privacy-conscious hiring workflow could look like end to end.

Current Approach

Working through the system architecture for collecting candidate signals, ranking candidates, and issuing time-limited coding assessments — treating data privacy as an explicit design constraint rather than an afterthought.

What I'm Testing

How to rank candidates fairly from public signals, and how to structure the assessment and review workflow so hiring managers get useful, privacy-respecting output.

Current Milestone

Working through the candidate evaluation and coding assessment workflow design.

Next Step

Refining the data privacy approach and the hiring-manager review experience.

Concepts & Technologies

Candidate Data CollectionCandidate RankingMachine LearningWorkflow AutomationCoding Assessment SystemsAutomated NotificationsData PrivacySystem Architecture
MORE DETAILS ON EXPERIMENT SOON

Where This Might Go

Future Experiments

Directions I expect to explore next — not projects yet, just where the curiosity is pointing.

LLM ApplicationsRAGAI AgentsComputer VisionML SystemsRoboticsMLOps

Right Now

What I'm Learning

Machine Learning

Classification, ensembles, evaluation

LLMs

Prompting, evals, application design

AI Agents

Tool use, multi-step workflows

Computer Vision

Applied CV for sensor-driven systems

Robotics

Embedded systems, real-time control

Data Engineering

Pipelines feeding ML systems

Cloud Infrastructure

Deployment, containers, CI/CD

Log

Lab Timeline

  1. AUG 2026

    Gesture-Controlled Drone

    Prototype development

  2. AUG 2026

    hushHush Recruiter

    AI recruitment automation

END_OF_LOG