Hi, I'm Wisam Damouny. Software Technical Leader at Qlik & SaaS founder. I architect high-performance enterprise data pipelines, self-hosted systems, and autonomous human-AI workflows.
I am a Software Technical Leader with 10+ years of experience engineering high-throughput enterprise data pipelines and self-hosted platforms. At Qlik, I lead SAP data integration teams and performance optimization across complex enterprise environments.
I hold a B.Sc. in Electrical & Computer Engineering from TU Darmstadt (Germany). Fluent in 4 languages—English, German, Hebrew, and Arabic—I specialize in building high-performance systems and environments where autonomous AI agents and human teams collaborate seamlessly.
Architecting secure, private software stacks that operate entirely on your infrastructure with low latency and zero vendor lock-in.
Building intuitive kanban and task systems where human developers and AI agents collaborate on real code bases.
First-class support for right-to-left UI layout dynamics, typography, and localized user experience.
From initial domain concepts to scalable APIs and responsive, micro-animated user interfaces.
TU Darmstadt (Darmstadt University of Technology, Germany) · 2009–2013
Specialized in low-level system design, multithreading, and ML systems. Final Project: Java ML Self-Learning Energy Monitoring System.
4 Languages across Global Tech Markets
Educational AI Tools & Mentorship
Building educational speech & literacy AI tools for children in Arabic and Hebrew. STEM volunteer mentor since 2004.
AI & Board
The project management board where humans and AI agents work together in real-time. Hebrew-native, self-hosted, and ultra-fast.
Domain Tools
Hebrew-first construction estimation & bill of quantities (BOQ) workspace platform for contractors and engineers.
Open Source
Open-source project management contributor — added comprehensive Hebrew translation, RTL layout, and modern dark themes.
Autonomy is a per-action decision, not a global setting — and the risk-classified approval gates, escalation matrices, and auditable human decisions that keep enterprise agents from acting with unearned authority.
Why RTL bugs live in the Unicode Bidi Algorithm, not your code — and the logical properties, Hebrew locale rules, and visual test discipline that make right-to-left products shippable.
Why raw LLM control flow does not scale in production — and the typed states, human-in-the-loop gates, checkpointing, and fallbacks that turn a stochastic agent into a deterministic, resumable machine.
The practical playbook for isolating agent tool execution — gVisor/Firecracker boundaries, seccomp syscall allowlists, egress control, and permission gates that contain an agent's blast radius.
Why production AI agents fail on observability — and the OpenTelemetry-based tracing, trajectory logs, and replay surface that turn "the agent seems fine" into proof.
Why most enterprise AI-agent initiatives die — and the sandboxing, authorization, and observability patterns that separate working autonomous systems from demo toys.
Stand up a Hermes autonomous AI agent in one click — connect a model provider, wire messaging channels, and know when to self-host instead.
The DIY path: a Nous Hermes 3 model as an autonomous tool-using agent on one CPU VPS — provisioning, Ollama serving, the agent loop, TLS hardening.
Key design principles for production multi-agent systems, tool execution sandboxes, state graph orchestration, and human-in-the-loop governance.
Architectural patterns from high-throughput enterprise data integration: backpressure handling, stateful streaming, and zero-allocation memory pools.
Why self-hosting developer tooling reduces cloud overhead by 70% while improving security compliance, latency, and full operational control.