Production Systems for
Real Workflow Problems.

Greater Angels Software is an engineering studio led by a Senior Principal Software Engineer, serving public agencies and public universities directly and teaming with prime contractors on state and local pursuits. We take the workflow problems your staff live with every day and solve them with supported production systems: integration for the vendor stacks you already own, engagement platforms your community actually uses, and auditable AI overlays where your staff keep the final say.

Discuss Teaming Explore Capabilities
California DGS Certified Small Business · Cert #2052888

SAM.gov Registered · NAICS 541511 · 541512 · 541519 · 518210 · Azusa, California

Augmentation,
Not Automation.

Public agencies and public universities run on high-stakes workflows and fractured vendor stacks, from 9-1-1 dispatch to permitting counters to campus student services. Technology there must protect the bandwidth of trained staff without ever compromising their oversight.

So we build two things, and build them well: the integration glue that unifies the legacy systems institutions already own into workflows they can actually run, and auditable, human-in-the-loop AI overlays where the AI drafts and your staff decides. We don't replace the human in the loop; we eliminate the bottlenecks around them.

The Greater Angels Advantage

Aerospace-Grade Rigor. Founded by engineering leadership with nearly a decade of experience building mission-critical systems in Fortune 100 defense and aerospace environments. We bring absolute CI/CD discipline and architectural resilience to civic deployments.

Deterministic AI Expertise. Led by data science experience in modeling complex, high-stakes scenarios (including NASA fire spread models), we specialize in rigorously constrained, hallucination-free ML frameworks that meet the State's strict Responsible AI standards.

Four Capabilities, Each With Proof

Agencies and primes face a false choice: pay national-consultancy rates for a rotating bench of junior staff, or stretch an already-thin internal team past its limits. We are the third option: a DGS Certified Small Business (SB) delivering locally accountable engineering, led and signed off by a Senior Principal Software Engineer, across four capability areas, and every claim below comes with working proof or an honest maturity label.

Legacy-System Integration Glue

Institutional software is a cobbled-together stack of vendor systems that were never designed to talk to each other. We build the connective layer nobody else wants to own: secure APIs and event-stream listeners for proprietary platforms, SSO and identity integration, reference-data sync, all delivered as containerized, failover-ready Docker/Kubernetes modules.

Proof: the Library Connect case study, unifying a catalog system, payment manager, and address verification behind one accessible intake flow. Read the case study.

Auditable Human-in-the-Loop AI Overlays

Retrieval-only search and conversational chat that answer only from your approved content, AI translation, and AI-assisted drafting, always with human review before anything commits and a full audit trail of every action. The AI drafts; your staff decides, and the code enforces it.

Proof: the Campus Wellness Passport’s retrieval-only guide answers only from Student Health Services content and deflects rather than inventing. Live on synthetic data. Read the case study.

AI Governance & Conformance Verification

Governance you can run, not a binder. We write an AI system's behavioral bounds as plain-language contracts: what it must always do, what it must never do, how well it must perform. We bind them to executable tests and deliver the evidence a procurement reviewer can actually read: conformance results, audit trails, human-approval gates, and SIMM 5305-F-shaped risk documentation. Not a certification. A test run with receipts.

Proof: our live Campus Wellness Passport demo enforces its own contracts in code. Its safety guardrails return before the model is ever called, answers come only from approved content, and the database can never claim a machine scored work a human actually scored. The evidence pack is published rather than promised: requirements, traceability, a WCAG 2.1 AA conformance report, and a completed HECVAT Lite. We also authored a full high-risk GenAI SIMM 5305-F for our own state RFI response. See it live.

Teaming-Ready Small Business

California DGS Certified Small Business #2052888, UEI PJK5UHM8PPR6, CAGE 21YS4, SAM.gov active, and headquartered in Los Angeles County, credit toward small-business participation goals and local-business preference programs. We team as severable, fixed-price work packages under your references, not body-shop staffing.

Proof: verifiable registrations, one click away on our Government page.

Ruthless Engineering Discipline

We operate with the rigor of federal defense programs. Our engineering teams work under strict, industry-leading methodologies to ensure zero compromises on public safety data security and system uptime.

Test-Driven Development (TDD)

Every line of code is proven by automated tests before it reaches production, ensuring absolute system reliability for the agency operations that depend on it.

Behavior-Driven Development (BDD)

Software logic is mapped directly to plain-English operational requirements, guaranteeing the technology precisely matches the agency's workflow intent.

Systems Architecture (SysML)

We model the complete IT ecosystem before writing code, ensuring our microservices integrate seamlessly with the enterprise systems your agency already runs.

Proof, Not Promises

Every claim on this page traces to a working artifact: an executable specification, a system diagram, or a running demo. The sample below comes from the SHS Digital Wellness Passport, built for CSU Bakersfield Student Health Services at the Cal Poly Digital Transformation Hub’s 2026 AI Summer Camp. The app is live at demo.greaterangelssoftware.com, and its engineering record is public, not on request: requirements, the traceability matrix, the 4+1 architecture views, a WCAG 2.1 AA conformance report written for CSU ATI review, and a completed HECVAT Lite for campus ITS security, all at docs.greaterangelssoftware.com. What this means for a campus department is laid out on our Higher Education page; the other case studies live on our Government page.

59 requirements, each traced to its tests 206 executable scenarios, run on every commit 86 architecture diagrams, published WCAG 2.1 AA conformance report HECVAT Lite, completed and cited 
Feature: QR check-in bound to the
         scanned event

  Background:
    Given I am authenticated and
          enrolled in the active
          challenge

  Scenario: Scanning a week's code
            credits exactly that week
    Given the event QR minted for
          "Vital Screenings"
    When I scan it
    Then the check-in is recorded
         against "Vital Screenings"
    And the confirmation shows week 2,
        not any other week

  Scenario: A stale poster does not
            credit a replacement task
    Given a signed token minted for a
          task that was later deleted
    And a different task now occupies
        its identifier in the active
        challenge
    When I scan the stale code
    Then I see "This code is no longer
         valid, ask the attendant"
    And no check-in is recorded
The requirement behind the gate shown at right, written as an executable specification and held verbatim against the requirements document. It runs as an automated test on every commit: the requirement is the test.
Sequence diagram: a scan of a stale event QR is rejected with HTTP 400 and writes no row, because the signed token carries the event it was minted for and not just a task id; a scan of the current code validates its signature and rotation window, resolves to the active challenge, and inserts one check-in under a uniqueness constraint that turns a second scan into a 409 rather than a double count
That same requirement as designed and as built: the signed token carries the event it was minted for, so a recycled identifier can’t quietly credit the wrong week, and a uniqueness constraint makes a second scan a rejection rather than a double count. The number staff report at the end of the challenge is one they can defend. Tap the diagram to open it full size.

Our Team

Primes and agencies buy from trusted teams, not anonymous landing pages. Here is who leads the work, and how we staff it.

Anthony Graca

Founder & Senior Principal Software Engineer

Nearly a decade of engineering on mission-critical federal aerospace programs, including flight software supporting the James Webb Space Telescope and Artemis. MS Computer Science candidate at Cal State Fullerton, researching measurable outcomes of apprenticeship-based engineering training. Every engagement is supervised, reviewed, and signed off by Anthony personally.

Matthew Graca

AI Engineer

Data science experience modeling complex, high-stakes scenarios, including NASA fire-spread prediction models. MS Computer Science candidate at Cal State LA, speaker at the ICDATA 2026 conference, with three papers currently in peer review. Matthew leads our rigorously constrained, fully auditable ML frameworks, designed to meet the State's Responsible AI standards without hallucination risk: the AI expertise behind every overlay we ship.

Celine Truong

Product Management Intern

Driving product discovery, validation, and go-to-market strategy. Recent B.S. Computer Science graduate from Cal State Fullerton, with experience conducting customer research, leading product validation experiments, roadmap planning, and cross-functional product execution. Celine helps bridge technical development with customer needs, ensuring every product decision is evidence-based and mission-focused.

Alejandro Silva

Software Engineer Intern

Alejandro is a software engineer and recent CSU Fullerton graduate specializing in backend architecture, data modeling, and REST APIs. He thrives on owning the entire development lifecycle by leveraging AI tooling to rapidly build robust, user-centric products like live recommendation engines and municipal library platforms.

The Engineering Team

Behind our leadership is a team of early-career engineers working under a Senior Principal Software Engineer’s direct supervision: the teaching-hospital model. Leadership reviews and signs off on every line, and nothing reaches an agency system without passing our test-driven pipeline. It is how we deliver production-grade software at new-grad wage rates rather than national-vendor billable rates, and it is the subject of our peer-reviewed engineering-education research.

Initiate a Teaming Discussion

We are actively teaming with prime contractors on California state and local procurements as the specialist sub for AI-overlay, integration, and compliance scope, and we work directly with public agencies and university departments on right-sized departmental purchases. Procurement officers can find our certifications, purchasing paths, and capability statement on our Government page; campus departments can find the accessibility and security artifacts a campus review will ask for on our Higher Education page.

Reach Us Directly

What to Include

  • Your agency or firm, and your role
  • The solicitation, RFI, or capability gap in question
  • Your intended purchasing path (teaming, SB/DVBE Option, departmental purchase)
  • Your timeline; we can turn a capability response around quickly