EdTech & Corporate Learning
EdTech Software Development - LMS, AI Tutors, Skills Assessment & Corporate L&D
EdTech is in the middle of an AI rewrite - the products that win will be the ones whose AI is pedagogically sound, not just demo-good. We build LMS, AI-tutoring, and corporate-learning platforms that respect learners and survive a procurement review.
What we hear from EdTech & Corporate Learning teams
- Off-the-shelf LMS that learners route around because the UX is from 2008
- AI-tutor products that confidently mis-explain a concept - and the K-12 buyer never trusts the product again
- Procurement reviews that surface FERPA / COPPA / GDPR-K gaps the team didn't know existed
- Corporate L&D platforms with content libraries nobody can find anything in
- Skills-assessment products with weak validity evidence - and an enterprise buyer who asks for it
- Accessibility (WCAG, Section 508) treated as a post-launch fix instead of a constraint
Regulation & compliance we work with
FERPA (US student data)
COPPA (US under-13)
GDPR + GDPR-K provisions (EU minors)
State student-data privacy laws (CA SOPIPA, NY Ed Law 2-d, IL SOPPA, etc.)
Section 508 + WCAG 2.2 AA (accessibility, mandatory for many public-sector buyers)
SOC 2 Type II
ESSA evidence tiers (where products make efficacy claims)
What we deliver
LMS replacement + custom-LMS builds with learner experiences people actually use
AI-tutoring platforms with retrieval grounding, scope guardrails, and pedagogical validity built in
Skills-assessment platforms with item-banking, validity-evidence workflows, and item-response-theory scoring
Corporate L&D platforms: content management, learning paths, microlearning, manager dashboards
Accessibility-first UI (WCAG 2.2 AA from sprint 1, not bolt-on)
Content-licensing workflows + DRM where rights-protected materials flow
Integration with SSO (SAML/OIDC), SIS (Clever, ClassLink, OneRoster), and HRIS for corporate L&D
FAQ
- How do you design AI tutors that don't mis-explain?
- Pedagogical scope, retrieval grounding, refusal behaviour, and a continuous evaluation harness. The tutor only answers from a known content corpus, refuses to wander outside scope, and is evaluated weekly against a labelled question bank with subject-matter-expert review of low-confidence outputs. We treat AI-tutor quality as a continuously measured product property, not a launch checkbox.
- Are you familiar with FERPA, COPPA, and state student-data privacy laws?
- Yes - and they're design constraints, not late additions. Student PII gets a separate data classification, retention policies are explicit, parental-consent workflows are first-class, and sub-processor inventories are kept current. For state-by-state laws, we maintain a compliance matrix that maps requirements to platform controls.
- Can you handle accessibility from the start (WCAG 2.2 AA / Section 508)?
- Yes - and it's a sprint-1 concern, not a launch-week fix. Components are built accessible from the design system up, automated accessibility tests run in CI, manual screen-reader testing happens before each release, and the VPAT is maintained as a living document. Many public-sector and enterprise buyers won't even start procurement without it.
- Do you build content / DRM workflows for licensed materials?
- Yes. Rights-protected content (textbooks, video courseware, assessments) needs licence-aware delivery, watermarking, expiry, and audit logs. We've built these workflows for both K-12 and higher-ed content publishers, and for corporate L&D platforms with third-party content libraries.
Who builds it
Our own team, working from our office in East Sarajevo, Bosnia and Herzegovina. The engineers on your scoping call are the ones who build it.
Meet the team