Aly Khaled

Aly Khaled

Building SafeDawa AI, a clinical safety intelligence platform for drug discovery.

Building

I'm building SafeDawa AI to predict clinical safety problems before drug candidates reach patients. The platform combines real-world safety data, knowledge graphs, and agentic reasoning to help drug discovery teams investigate risk faster and with more depth.

40+

tools shipped in a dynamic registry

1K+

jobs orchestrated each month

20+

drug discovery projects supported

99.9%

job status accuracy in production analytics

I'm Aly Khaled, an engineer from Egypt with a background in biomedical and software engineering. Over the years, I worked across healthcare platforms, machine learning infrastructure, and scientific software, building systems for clinical environments, research teams, and data-heavy workflows.

Early in my career, I worked on products that had to be useful under real constraints. I helped build systems that supported more than 100,000 users in clinical settings, built ML tooling that cut model size by 50% and training time by 70%, and learned how to ship software that had to perform, scale, and actually get used.

At Proteinea, I worked on scientific and AI infrastructure for antibody engineering and computational biology. That meant building compute platforms, tool registries, job orchestration systems, monitoring pipelines, analytics dashboards, and research products used in large-scale experimentation.

Proteinea shaped how I think about software. I like technically deep products where infrastructure, data, AI, and product design all matter at once. I care about systems that are operationally serious, readable to their users, and strong enough to become part of how teams make decisions.

That path led me to SafeDawa AI, the company I'm building now. SafeDawa is an AI-powered clinical safety intelligence platform for drug discovery. The goal is to predict clinical safety problems before drug candidates reach patients, focusing on outcomes like black box warnings, market withdrawals, organ-specific injury patterns, and time-to-safety-action.

Under the hood, we're building the Clinical Outcome Knowledge Graph, which connects 16.5 million FDA FAERS adverse event case reports with structural similarity networks, ChEMBL target relationships, regulatory action history, and mechanistic pathway models from AOP-Wiki.

On top of that graph, we're building ToxDetective, an agentic AI safety scientist that investigates compound safety profiles the way a senior toxicologist would and turns that work into actionable reports with experimental next steps.

What drives me is simple: I like building systems that make ambitious people more powerful. SafeDawa is the clearest version of that idea I've worked on so far, and the kind of company I want to build is straightforward: serious software, real scientific leverage, and products that compress weeks of expert investigation into minutes.

  • Turn complex scientific data into decisions teams can actually act on.
  • Compress expert workflows from days into minutes without losing rigor.
  • Build software that is technically serious, operationally reliable, and directly useful.

If you're working on drug discovery, clinical safety, scientific software, or AI systems that need real-world rigor, we should talk.

I'm building SafeDawa at the intersection of clinical data, knowledge graphs , and agentic AI. I'm always interested in conversations with founders, researchers, operators, and partners working on serious problems. Email is the fastest path.