Research dossier / Melbourne, Florida

Building AI systems that know what they know.

I’m Muntaser Syed. I research and build systems for trustworthy AI, agent infrastructure, edge intelligence, semantic data, and efficient machine learning.

Explore the research Browse the software

Portrait of Muntaser Syed — AI systems researchOpen source

14Verified papers & preprints

22Active Python packages

3Language ecosystems

2Devpost portfolios

Research map / 01

Research areas.

My work covers epistemic trust, compact data representation, edge computation, and practical AI infrastructure. The links below connect each area with its papers and software.

Trustworthy & semantic AI

Machine-readable confidence, provenance, temporal validity, compliance, and knowledge verification.

Agent systems

Governed execution, context budgets, deliberative simulation, and reproducible agent infrastructure.

Edge AI

Intelligence that fits constrained devices, networks, and real-world sensor systems.

Efficient ML

Physics-grounded energy measurement, compact architectures, and frugal model deployment.

Health & embodied intelligence

Wearable sensing, medical imaging, speech systems, robotics, and privacy-aware health AI.

Data systems

Hash-augmented structures and compact interchange for high-integrity machine knowledge.

Selected work / 02

Recent publications.

A selection from the verified catalog. The research page reconciles DOI records, proceedings, arXiv, and author profiles into one deduplicated bibliography.

2026

Preprint

The Environmental Cost of Digital Sovereignty

Water, energy, and emissions impacts of sovereign AI infrastructure in the Global South.

arXiv

2026

FLAIRS-39

JSON-LD 1.2 and Beyond

Extensions for machine learning data exchange, uncertainty, provenance, temporal scope, and security.

DOI Code

2026

FLAIRS-39

TOML: Transistor Operations for Machine Learning

A physics-grounded framework for measuring machine-learning energy efficiency beyond FLOPs.

DOI Code

2026

FLAIRS-39

Real-Time Neck Posture Classification Using a Lightweight Wearable IMU Pendant

Single-IMU posture classification designed for continuous use on constrained wearable hardware.

DOI

View the complete publication record

Open source / 03

Major open-source projects.

The software page groups related packages and ports. It documents the problem, the implementation, and the available Python, Rust, or TypeScript release.

Explore the software families

TRUST / SEMANTICS

JSON-LD Ex

Confidence, provenance, temporal validity, privacy metadata, and secure linked data for AI systems.

PythonTypeScript

AGENT INFRASTRUCTURE

Tokenmaster + ctxmaster

Context-budget metering, prediction, visualization, compaction, and handoff decisions for LLM applications.

PythonRustnpm

DATA STRUCTURES

Hashrope

Balanced hash-augmented ropes for sub-linear comparison, routing, repetition encoding, and bioinformatics.

PythonRust

Recognition / 04

Recognition and public profiles.

Named one of Major League Hacking’s Top 50 hackers.

MLH Top 50, Class of 2020

View the MLH Top 50 profile

Primary Devpost portfolio Secondary Devpost portfolio LinkedIn Florida Tech feature

Collaborate / 05

Contact.

Email me about research collaboration, open-source work, technical talks, or applied AI projects.

muntaser@ieee.org