The company centers on four product lines: Alfred, the company brain that executes; Cejel, a trust certificate for code; Egbert, our B2B Fintech 3.0 infrastructure platform; and Therasyn, governance and on-prem infrastructure for clinical AI.
Applied AI products for studios, markets, and regulated health systems.
Four focused surfaces, each built around a concrete operating environment.
The operator OS that runs a studio — or a whole portfolio — on AI, safely: hard isolation between products, governed execution that proposes but never ships on its own, verified memory, and a trust certificate for everything the AI builds. Cloud or fully on-prem.
A free, offline CLI that scores how trustworthy any codebase is, in any language — tests, secrets, isolation, CI and audit discipline — folding your existing scanners (Snyk, Semgrep, OpenSSF Scorecard) into one portable, shareable certificate. Especially useful for AI-written code — and any code you didn't write yourself, like an acquisition or a legacy system — where trust can't be eyeballed.
Infrastructure for validating trading strategies and moving them from research to governed paper pilots and production — with explicit risk controls, full attribution, and human oversight. Includes Edred, an autonomous strategy and model-health agent that proposes tuning as human-reviewed changes — propose-only, never trading on its own.
Governance and deployment infrastructure for clinical AI in environments where cloud-only isn't an option — academic medical centres, regulated health systems, and pharma research. BAA-bound by design, on-prem capable from day one. Built so health systems can run AI inside their own perimeter, with audit, PHI controls, and required sign-off enforced by construction.
Barg Labs builds applied AI products around real operating constraints: who is using the system, where the data lives, what has to be reviewed, and what must remain under human control.
The work is product-led rather than demo-led. Each surface is designed for a specific audience and an environment where reliability, legibility, and iteration matter.
Barg Labs is led by Houman Azimi-Nejadi, a software engineer and founder with more than 25 years of experience building complex software systems across multiple domains.
The company works closely with expert collaborators across product, clinical, market, and web systems as each product line moves from focused surface to live deployment.
Alfred is the company brain that executes — the operator OS that runs a company, or a whole portfolio, on AI, safely. As agents do more of the building and operating, the blocker is trust; Alfred is the layer that makes it safe.
It runs on hard per-product isolation (one product's secrets never touch another's), governed execution (agents open reviewed changes and run operations but never merge, deploy, or leak, with a full audit trail), and verified memory that only trusts what a real outcome proved. Cejel, its trust certificate, verifies what the AI builds before it ships.
Cloud or fully on-prem — including on a local model — so a regulated studio can run everything inside its own walls.
Live operator surface available at alfred.barglabs.ai.
Cejel is a trust certificate for code — any codebase, in any language, whoever or whatever wrote it. It is a free, offline command-line tool that scores the engineering signals that tell you whether to trust a repository - tests, secret hygiene, isolation, claim-vs-reality, and CI and audit discipline - and prints a portable certificate.
Rather than competing with the scanner you already run, Cejel aggregates them. Pipe in SARIF from Snyk, Semgrep, or CodeQL, plus OpenSSF Scorecard, and get one shareable, rubric-scored certificate over all of them, with every contributing tool attributed. It is especially useful for AI-written code, and for any code you didn't write yourself - an acquisition, or an inherited legacy system - exactly when trust can't be eyeballed.
Open-source and source-available. Runs fully offline, with no signup and no model call. A watermarked, air-gapped build runs fully on-prem for regulated teams.
Egbert is Barg Labs' B2B Fintech 3.0 infrastructure platform for validating trading strategies and moving them from research to governed paper pilots and production, with explicit risk controls and human oversight.
It is designed for real execution, monitoring, and controlled rollout, so strategy work can move from analysis into action without becoming an opaque black box.
Egbert includes Edred, an autonomous strategy and model-health agent: it reads live model and trading signals, proposes parameter tuning and re-weighting as human-reviewed changes, and flags model-health issues — propose-only, fully audited, and gated by human approval. It never trades or changes strategy on its own.
Therasyn is governance and on-prem infrastructure for clinical AI in environments where cloud-only isn't viable - academic medical centres, hospital systems, pharma clinical research, and any setting bound by BAA + state law + sovereign-data requirements.
It is BAA-bound and on-prem capable from day one. Health systems run AI inside their own perimeter - with audit trails, PHI controls, and required sign-off enforced by construction - over multimodal clinical records (genomics, pathology, radiology, notes), without that data ever leaving the institution.
For Therasyn the binding constraint is never "does the technology work" - it is "can the technology live where the data has to live." Therasyn is built for exactly that constraint.
For product conversations, partnerships, or early access, contact the team directly.
team@barglabs.ai