AI systems engineering,
done right.
Modulus Labs exists because we saw the same pattern too many times.
Talented engineering teams building AI prototypes that worked in demos but collapsed in production. Not because the models were bad, but because the systems around them — the evals, the monitoring, the fallback logic, the security layers — were treated as afterthoughts.
We knew we could do better. AI capabilities were advancing rapidly, but most teams were barely scratching the surface of what was possible. We saw technologies with limitless potential and made a choice: use them to the fullest extent of their capabilities, push for the best possible outcomes, and build the engineering systems to make that real.
Today, Modulus Labs is a focused AI systems engineering company. We do not build AI research. We do not train foundation models. We build the production systems that turn AI capabilities into reliable business outcomes — and we hold ourselves to the same engineering standards we would expect from the teams we work with.
Who we are
Mujtaba Raza
Co-founder & CEO
Mujtaba leads Modulus Labs — setting company strategy, driving technical direction, and architecting the AI systems the company ships. A lawyer by training, he moved into the energy sector as founder of Solar Citizen, building real-time monitoring and predictive intelligence platforms for solar operations at scale. That experience — operating at the intersection of complex systems, data infrastructure, and business outcomes — became the foundation for Modulus Labs. He now builds across RAG pipelines, LLM applications, multi-agent ecosystems, IoT, and full-stack SaaS, bringing a rare combination of legal reasoning, engineering depth, and commercial instinct to every engagement.
Amaan Saigol
Co-founder & COO
Amaan leads business operations, client partnerships, and market expansion at Modulus Labs. With a background in business development and technology ventures, he bridges the gap between technical capability and commercial impact — ensuring every engagement delivers measurable business outcomes. He drives the company’s growth across European and Middle Eastern markets.
What we believe
Outcomes over activity
We measure our work by what ships and what it achieves — not by hours logged or lines written. Every engagement has defined success criteria, and we hold ourselves to them.
Honesty as a service
We will tell you when a simpler solution is better. We will tell you when AI is not the right tool. Our job is to solve your problem, not to sell you more AI.
Craft matters
Clean code, thoughtful architecture, comprehensive tests. We take pride in systems that are a pleasure to maintain — not just systems that work today.
Transparency by default
You see everything: our architecture decisions, our tradeoff reasoning, our progress, our mistakes. No black boxes. No surprises at the end of a sprint.
Your team gets better
We transfer knowledge as we build. Documentation, pairing sessions, architecture walkthroughs. When we leave, your team can maintain and extend what we built.
What makes our engineering different
Most AI consultancies optimize for impressive demos. We optimize for the six months after deployment — when the real work begins. Models drift, data distributions shift, edge cases multiply, and costs compound. The systems we build are designed for this reality.
Every project starts with measurement. Before we write a single line of application code, we define success criteria and build the evaluation framework to track them. This is not overhead — it is the foundation that makes everything else possible. When you can measure quality automatically, you can iterate with confidence, catch regressions before users do, and make data-driven decisions about what to improve next.
We treat AI components with the same rigor as critical infrastructure. That means comprehensive error handling, fallback chains for when models fail, circuit breakers for external dependencies, and health checks that surface problems before they become incidents. Our systems degrade gracefully instead of failing catastrophically.
We do not subscribe to "move fast and break things." We move deliberately and build things that last. Speed comes from making the right decisions early — not from cutting corners that create technical debt.
Specific practices, not slogans
Every team claims quality. Here is what ours looks like in practice.
Evaluation-first development
We build the eval suite before we build the feature. Every AI component has measurable quality criteria that are tested automatically on every change.
Code review on everything
Every line of code is reviewed by a second engineer. No exceptions for prototypes, no shortcuts for urgent work. This is where we catch the mistakes that matter.
Production monitoring
Every system ships with logging, metrics, and alerting. We monitor model quality, latency, cost, and error rates — not just uptime.
Security audits
AI-specific security review on every project. Prompt injection testing, output validation, PII leak detection, and access control verification.
Documentation as deliverable
Architecture decision records, operational runbooks, and API documentation are not afterthoughts. They ship with the code because they are part of the product.