We build product-grade AI and run it as managed programmes — verifiable, sovereign, and explainable — for organizations in cybersecurity, healthcare, education, and intelligent automation. Research-led, delivery-proven across MedTech, FinTech, and EdTech.
We deliver AI as operational programmes for high-trust environments, and operate our own products where the work matures into something reusable.
Managed research programmes for data-heavy organizations that need verifiable insight generation and prediction pipelines.
Operational AI workflows for anomaly analysis, risk scoring, and explainable security operations in regulated settings.
Human-led personalized learning systems blending generative content, pedagogical control, and measurable outcomes.
AI cybersecurity platform with deployable agents for audit, network analysis, and secure browsing.
Visit clara.stm.aiAI-driven personalized education product for children with digital and printable learning experiences.
Visit llmentor.stm.aiResponsible computation paired with deep domain collaboration — so advanced models become trusted operational systems, not isolated experiments.
Domain experts and stakeholders stay inside every review, governance step, and final decision, shaping each iteration through continuous feedback — never replaced by the model.
We design for on-prem and controlled-hybrid environments, so data custody, policy, compliance, and execution control stay inside your boundary.
Every output is tied to its data lineage, rationale, and provenance, with audit-ready reporting that makes results inspectable and verifiable.
Dedicated cross-functional teams own delivery from roadmap to operations, with shared accountability and measurable milestones — outcomes, not advisory billing.
Peer-reviewed work and field notes — the body of research behind everything above.
How a phone became an MHRA-certified Class I medical device that assesses Parkinson's motor symptoms at home — using deep learning to guarantee data quality and cut a 25-minute test to under 4.
A prospective, dual-site study testing whether a 16-item smartphone assessment can predict the motor scores three blinded neurologists give — and what rigorous validation means for digital measures in Parkinson's trials.
An open-source Python toolkit that standardises how Parkinson's symptoms are measured from wearables and smartphones — turning raw sensor data into transparent, comparable biomarkers and clinical scores.
How AI STM Learning and SmartClover are building an evidence-driven NIS2 compliance and resilience accelerator for SMEs in essential and important sectors.
stm.ai combines product execution with academic and applied R&D experience across regulated, educational, and data-intensive sectors.
Parkinson's disease analytics, smartphone-based clinical assessment, and health data science toolkits.
cloudUPDRS, PDkit, peer-reviewed clinical papers
Transfer-learning and forecasting workflows for financial applications and trading knowledge systems.
Transfer learning publication, market intelligence delivery
Personalized AI learning products and content pipelines designed for children, educators, and families.
LLMentor product development and programme execution
Academic researcher and lecturer background with practical education design and training-focused delivery.
Birkbeck profile, Fundamentals of Teaching certification
Explainable AI systems for security operations, risk analysis, and governance-oriented deployments.
CLARA platform and secure deployment architecture
Interdisciplinary experiments combining cognitive science, AI, and public engagement design.
Me, Human collaboration with Science Museum
Selected past builds beyond our clinical research. Our Parkinson's work (cloudUPDRS, PDKit) now lives in the research section.
Intelligent commerce platform research with Romanian ASR, NLP product reasoning, and visual recognition.
LLM-driven extraction and product intelligence with crawler-assisted data pipelines and vector workflows.
Public science collaboration on human cognition and behavior through live experiments and research outreach.
Custom Python and SQL forecasting stack for structured market knowledge capture and decision support.
For partnerships, proposals, pilots, or case-study details, email c@stm.ai.
Founded by Cosmin Stamate, stm.ai combines product engineering, machine learning research, and proposal-ready technical planning. Legal entity: AI STM Learning SRL.
We collect only data sent by email for partnership communication. Data controller: AI STM Learning SRL. You can request access, correction, or deletion at privacy@stm.ai.