Less hype. More code.

"Humans were creating unmaintainable spaghetti code long before AI existed; blaming AI is just unfair."

Anyone with an LLM prompt can generate a generic CRUD app in five minutes. That's not interesting.

What AI cannot do is look at a broken, complex problem from an unexpected angle, connect two unrelated domains, and architect a system that shouldn't theoretically work—until it does.

I don't obsess over memorizing boilerplate; I use AI and modern tooling relentlessly to compress execution time so I can focus 100% on creative system design, unconventional problem-solving, and deep architecture (mostly in Rust, Python, and TypeScript). I bring the blueprint, the constraints, and the lateral thinking; the code is just the final medium.

Instead of building another wrapper, I like tackling problems that require rethinking the mechanics from the ground up:

• TypesetOK — Rethinking Desktop Publishing from Scratch
  The Creative Leap: Most developers avoid text shaping and layout engines because they are notoriously painful—especially when combining multi-stream flows with complex Right-to-Left (RTL) scripts and Hebrew vocalization (nikkud). Designing a reproducible PDF and typesetting engine in Rust required treating typography not as UI, but as a strict mathematical constraint problem.

• HDDE Architecture — Zero-Allocation Decision Engine Research
  The Creative Leap: How do you run thousands of autonomous agents with deterministic predictability and zero memory allocations during runtime? By fusing Utility AI, hierarchical command propagation, and cache-coherent ECS into a single pipeline. A pure exercise in lateral performance thinking.

• Otzaria Semantic Search — Deep Context over Ancient & Complex Corpora
  The Creative Leap: Standard keyword search fails when dealing with dense, multi-layered textual corpora where meaning is implicit rather than explicit. This engine bridges modern high-density vector embeddings with fast Rust-based retrieval to find ideas, not just matching strings.

I’m always drawn to unconventional ideas—projects that sit at the intersection of hard algorithmic constraints, deep domain puzzles, and creative product thinking. If you have strong engineering execution or a sharp domain problem and need a creative partner to help architect, brainstorm, or co-build the solution:

• Web / Portfolio: amlaach.github.io/personal-site (reach out directly via the site)
• Community: Catch me on the Bina Top (בינה טופ) forum — just send a private chat to 'המלאך'.