Danial Asgari

Danial Asgari

Computer Engineering Student · AI Systems & Reliability

Focused on designing intelligent systems that remain stable, adaptive, and reliable under uncertainty and real-world operational constraints.

My interests span AI systems architecture, self-regulating computational models, machine learning infrastructure, and system-level reliability engineering.

AI Systems
System Reliability
ML Infrastructure
Data Engineering

Experience

IT Specialist

Tizbam Educational Company

Nov 2024 – Present

Maintain IT infrastructure for e-learning platforms, resolving 10–15 software and hardware issues weekly. Automated data workflows and log analysis with Python scripting, and collaborate with educators and developers to support platform reliability — practical, daily exposure to how systems behave and fail in production.

Alongside the technical role, I volunteer as an organizer for Tizbam's preparatory exam events — coordinating exam-day logistics and guiding students through the process. Working as part of the organizing team under time pressure taught me operational coordination and responsibility for other people's experience, not just systems.

Junior IT Specialist

Tizbam Educational Company

Oct 2022 – Nov 2024

Supported the IT Specialist with day-to-day infrastructure and technical support tasks, gaining hands-on experience with company systems and internal data workflows — the stretch where I moved from following instructions to owning problems myself.

IT Intern

Tizbam Educational Company

Aug 2022 – Sep 2022

Learned SQL and database fundamentals while assisting with basic technical support tasks under senior staff supervision — my first real exposure to how a company's systems actually run, started the same month I began university.

Projects

Selected engineering work. Each project is an applied step toward the same question: how do you build systems that understand and manage their own behavior?

Self-Regulating ML Inference Service

View on GitHub →

A containerized machine-learning service that does more than serve predictions — it monitors its own performance, evaluates its prediction confidence, and regulates its own behavior. When confidence drops below threshold, it returns a safe "uncertain" response and flags itself for retraining through a self-assessing health endpoint. An applied realization of the principles in my SRCA research: a system that recognizes the edges of its own competence.

Python FastAPI scikit-learn Docker

Transcript Cleaner

View on GitHub →

A fully local, AI-powered transcription pipeline combining Whisper speech-to-text with a locally-run language model to clean and refine raw transcripts — entirely offline, with no external API dependency. Built end to end: backend service, web interface, and conservative LLM prompting designed to correct without rewriting the speaker's meaning.

Python Whisper Ollama FastAPI

RL Autonomous Racing Agent

View on GitHub →

A reinforcement-learning agent trained with PPO to drive autonomously in a simulated environment, extending an interest I first developed through the AWS DeepRacer League into a fully open, self-built implementation. Average training reward improved from approximately −58 to +121 over 100,000 training steps — a system learning, visibly, from its own mistakes.

Python Stable-Baselines3 Gymnasium Reinforcement Learning

Direction

My long-term goal is to contribute to the development of intelligent systems that can adapt, learn, and operate reliably in complex environments.

I am particularly drawn to the intersection of machine learning, system reliability, and autonomous decision-making — the space where a system's intelligence matters less than whether it can be trusted. Years of maintaining real infrastructure taught me that the hard problem is rarely making something work; it is making something keep working when conditions change.

Whether through research, engineering, or collaboration with leading institutions, I aim to help build technologies that are not only intelligent, but trustworthy and resilient. I treat every project, research effort, and learning opportunity as a step toward that goal.

AI Systems & Reliability

Designing systems capable of adaptive behavior while maintaining stability, fault tolerance, and operational trustworthiness.

Self-Regulating Architectures

Exploring feedback-driven computational structures that support autonomous adaptation, coherence maintenance, and intelligent regulation.

Data & Infrastructure Engineering

Scalable data systems, machine learning infrastructure, and system-level engineering for real-world intelligent applications.

Research

Flagship Research

Self-Regulating Cognitive Architecture (SRCA)

Independent research preprint (single author) — indexed on PhilPapers.

SRCA proposes a self-regulating model for artificial awareness based on multi-domain cognitive organization and closed-loop coherence maintenance. The framework integrates perspectives from Integrated Information Theory, Global Workspace Theory, and predictive processing into a unified, testable architecture for systems that regulate their own internal states.

Key Contributions

•

Four-domain cognitive architecture (E-U-P-C structure)

•

Self-regulation and feedback-loop mechanisms

•

Formal distinction between awareness and computation

•

Testable architectural implications for intelligent systems

Read the Preprint on PhilPapers →

Applied companion project: Self-Regulating ML Inference Service →

Research Interests

Self-healing systems using reinforcement learning

Fault prediction in distributed systems

AI-driven reliability and failure detection in complex infrastructure

Scalable architectures for intelligent monitoring in cloud and edge computing

Beyond Engineering

Before systems, there was competition — in science and in sport. Both shaped how I approach problems: preparation, discipline, and performing under pressure.

Iranian Physics Olympiad

Participated in the Iranian Physics Olympiad in 10th grade — an early, formative encounter with rigorous problem-solving beyond the standard curriculum.

Competitive Sports

Regional and provincial medalist in chess, swimming, and basketball — strategy, individual endurance, and team coordination, respectively.

Contact

Open to research collaboration, technical discussions, and opportunities in AI systems and data engineering.

✉️

Email

danialasgari1383@gmail.com

🐙

GitHub

github.com/daniasg

💼

LinkedIn

linkedin.com/in/danial-asgari

Download Academic CV (PDF) ↓