Research at Octargus

Oculomotor Intelligence

Computer vision for quantitative eye-movement analysis.

Exploring whether commodity cameras and modern computer vision can quantify subtle patterns in human movement and behavior. Initial research investigates webcam-based eye movement and oculomotor measurement — fixation, smooth pursuit, saccadic movement, gaze stability, and related quantitative features.

Technical areas

• Eye & gaze tracking
• Computer vision
• Facial landmarks
• Motion analysis
• Time-series analysis
• Multimodal signal processing
• ML pattern recognition

Exploring AI-enabled measurement of oculomotor behavior using consumer hardware.

Heme AI

AI-assisted analysis of blood cell morphology.

Exploring computer-vision and machine-learning systems for identifying, classifying, and characterizing cells in hematology microscopy images — extracting structured information from blood-smear images and other hematology imaging data.

Technical areas

• Cell detection
• Cell segmentation
• Cell classification
• Morphology analysis
• WBC differentiation
• RBC morphology
• Platelet identification
• Image quality assessment
• Vision model evaluation
• Human-in-the-loop review
• Structured reporting

Research into AI-assisted hematology image analysis and cell morphology classification.

Audio Intelligence

Extracting structured information from sound.

Exploring how machine learning, signal processing, and multimodal AI can extract quantitative features and patterns from speech and other audio signals — transforming unstructured sound into measurable features and structured information.

Technical areas

• Speech analysis
• Acoustic features
• Voice characteristics
• Prosody
• Pitch & frequency
• Speech timing & cadence
• Pause patterns
• Spectral analysis
• Audio embeddings
• Temporal patterns
• Audio classification
• Multimodal analysis

Exploring how AI can quantify patterns in speech and audio signals.

Scientific AI

AI systems that can work with scientific knowledge.

Exploring AI systems that combine foundation models with scientific literature, experimental data, protocols, structured databases, and domain expertise.

Technical areas

• Scientific knowledge systems
• Literature intelligence
• Experimental data interpretation
• Protocol understanding
• Scientific reasoning
• Retrieval-augmented systems
• Domain-specific AI
• Research copilots
• Multimodal scientific data

Intelligent Workflows

From AI assistants to systems that can act.

Investigating how AI agents can interact with software, data, knowledge systems, APIs, and human experts to perform complex research and knowledge-intensive workflows.

Technical areas

• AI agents
• Tool use
• Workflow orchestration
• Multi-agent systems
• Knowledge retrieval
• Human-in-the-loop
• Automated analysis
• Software interaction
• Agent evaluation
• Reliability & observability