AI Engineer & Founding Product Architect
PhD Candidate in Reinforcement Learning for Cyber-Physical Systems

I build AI-native systems that bridge reinforcement learning research and production-grade software.
My work focuses on reproducibility, system design, and high-stakes decision-making environments, where correctness, evaluation, and long-term maintainability matter.

I’m particularly interested in turning research-grade AI into robust, scalable products.

• AI-native collaboration tools for scientific and technical workflows
• Human–AI interaction patterns for complex authoring and decision-making tasks
• Reproducible experimentation pipelines for reinforcement learning systems

Founding Product Architect
🔗 Repo: (private / in development)

• Architecting a real-time, AI-assisted collaborative platform for scientific and technical writing
• Focus on structured documents, reactive data models, and human–AI co-authoring workflows
• Emphasis on reproducible research artifacts and long-term maintainability
• Tech: TypeScript, TanStack, Convex / PostgreSQL

Founding Mobile Engineer
🔗 Repo: (private)

• Built an AI-driven conversational mobile app for spoken language proficiency
• Owned the full product lifecycle: UX design, AI interaction patterns, subscriptions, and iteration
• Refined features based on user feedback and engagement metrics
• Tech: React Native (Expo), TypeScript, GenAI APIs

Founding Engineer
🔗 Repo: (maintenance mode)

• Designed a full-stack system for real-time project tracking and data synchronization
• Focus on data consistency, scalability, and operational clarity across distributed sites
• Tech: Next.js, TypeScript, Supabase

Alongside product work, I conduct research in reinforcement learning for cyber-physical systems, with applications to power grid resilience and security.

• Published in PeerJ Computer Science
• Focus on evaluation under adversarial and constrained environments
• Open-source experimental artifacts available for reproducibility

Paper: https://doi.org/10.7717/peerj-cs.3358
Artifacts: https://osf.io/gcw7x/

AI & Research

• Reinforcement Learning (PPO, TRPO, A2C)
• Simulation-based evaluation & ablation studies
• Reproducible experimentation pipelines

Systems & Product

• Distributed systems & data modeling
• Cloud-native architectures (Google Cloud – Certified)
• Production-first engineering and CI/CD

Stack

• Python, TypeScript, SQL
• PyTorch, React, Next.js, TanStack, React Native
• Docker, Git, Supabase, Convex, Grid2Op

• I favor simple systems that scale over clever abstractions
• I care deeply about reproducibility and long-term maintainability
• I optimize for clarity — in code, data, and interfaces

• Website: https://assembensalah.com
• GitHub: https://github.com/assemsohaib
• LinkedIn: https://www.linkedin.com/in/assem-bensalah

I pin repositories that best represent:

• system design decisions,
• research-to-production transitions,
• and maintainable, real-world codebases.