# Sparky*

> Sparky* (legal name Sparky solution d.o.o.) is a Croatian deep-tech company, founded in 2019, that connects state-of-the-art science with modern industry. A 15-person team of data scientists, engineers, and project and product people turns data and AI into working products, research results and reusable platforms for clients in Europe, the UK and the USA. 40+ projects delivered (45 published case studies) and 50+ scientific publications.

Sparky* is written with an asterisk after the name ("Sparky*"); "Sparky" and "Sparky Science" refer to the same company. Website: https://sparky.science. Contact: info@sparky.science. Careers: careers@sparky.science.

Key facts:
- Legal entity: Sparky solution d.o.o., Slavonska 9 (2nd floor), 10430 Samobor, Croatia (EU). EU VAT ID HR75796160484. EU PIC 885019077.
- Founded: 2019. General Manager and founder: Živko Krstić.
- Team: 15 employees, all in Croatia (main R&D hub at the University of Split, innovation hub in Zagreb, legal headquarters in Samobor, STEAM education hub on Lastovo).
- New York commercial hub (447 Broadway, New York) runs through the separate US entity Sparky solution LLC.
- Three business streams: (1) data and AI development and consulting, (2) research (EU-funded projects and internal R&D), (3) data and AI product development.
- Member of NVIDIA Inception.
- Tagline: "We turn bold ideas into working innovation*".

## Main pages

- [About](https://sparky.science/about/): who Sparky* is, how the team works, partners and offices
- [Services](https://sparky.science/services/): data science and AI/ML, data engineering, DevOps and MLOps, security, visualisation and apps, data and AI consulting
- [Projects](https://sparky.science/projects/): 45 case studies in 8 problem families across 22 industries
- [Products](https://sparky.science/products/): Sparky* Platform (Nexus, Atlas, Synapse, Nova, Pulse), Robotics, Space, No-Refunds, IoT, Pulse Sport
- [Research](https://sparky.science/research/): EU projects (SMART Researchers, ESG4PMChange+, HealthChain), publications, open datasets, academic partners
- [Careers](https://sparky.science/careers/): working at Sparky*, internships and student placements
- [News](https://sparky.science/news/): EU research, events, press, health, data and AI
- [Contact](https://sparky.science/contact/): hubs in Croatia, New York and Milan (partner), contact form

## Services

- Data science and AI/ML: predictive models, NLP, classification, forecasting, anomaly detection, GenAI, RAG, multi-agent systems
- Data engineering: ingestion pipelines, data lakes and lakehouses, streaming, data quality, integration (Databricks, Spark, cloud)
- DevOps and MLOps: infrastructure as code, CI/CD, model deployment and monitoring, cloud and on-premise
- Security: security reviews, access control, privacy and compliance checks; GDPR-ready by design
- Visualisation and apps: dashboards, analytics front ends, full-stack JavaScript applications
- Data and AI consulting: use-case scoping, data readiness checks, architecture assessments, stalled-pilot reviews, proof-of-concept design
- Every project includes security and compliance checks, documentation and handover, and knowledge-transfer sessions.

## Products

- [Sparky* Platform](https://sparky.science/products/#platform): five accelerators. Nexus (universal data ingestion and entity extraction across 20+ file types), Atlas (knowledge graphs and Graph RAG), Synapse (multi-agent orchestration and LLM routing across providers), Nova (chatbot, dashboards, workflows, docs, sheets, presentations, canvas, code, animation), Pulse (IoT and wearables: Pulse Health, Pulse Motion). Cloud, on-premise, hybrid or air-gapped. Status: in use.
- [Robotics](https://sparky.science/products/#robotics): Sparky*'s own AI layer on Pollen Robotics' open-source Reachy Mini and Microduck robots, for education, patient wellbeing and new roles. Status: R&D.
- [Space](https://sparky.science/products/#space): in-house AI decision system for real-time satellite image analysis. Status: R&D.
- [No-Refunds](https://sparky.science/products/#no-refunds): Sparky*'s first game, free to play from May 2027; foundation for game AI and a 3D engine for digital twins and simulations. Status: in development.
- [IoT](https://sparky.science/products/#iot): environmental monitoring of oceans, forests, manufacturing and camera sensors. Status: deployed.
- [Pulse Sport](https://sparky.science/products/#pulse-sport): medical-grade wearable sensors and AI for player performance; extension of HealthChain. Status: pilot.

## Case studies by problem family

- [Data platforms](https://sparky.science/portfolio-category/data-platforms/): data lakes and ingestion for finance, telecoms, energy, media and environmental agencies
- [Real-time intelligence](https://sparky.science/portfolio-category/real-time-intelligence/): streaming analytics, anomaly detection, wearable health dashboards, speech-to-text
- [Web intelligence](https://sparky.science/portfolio-category/web-intelligence/): reputation, media and market intelligence from online content
- [Document AI](https://sparky.science/portfolio-category/document-ai/): GenAI copilots, RAG assistants, document classification and extraction
- [Forecasting](https://sparky.science/portfolio-category/forecasting/): clinical decision support, valuation, churn, sports analytics, asset health
- [Operations](https://sparky.science/portfolio-category/operations/): route planning, fleet risk, smart waste, supply chain
- [Customer AI](https://sparky.science/portfolio-category/customer-ai/): recommendations, loyalty, CRM, matching, learning platforms
- [Immersive](https://sparky.science/portfolio-category/immersive/): virtual reality for energy grids and customised products

## Research

- [SMART Researchers](https://smartresearchers.eu/) (Horizon Europe, 2025 to 2028): industry partner supporting early-career researchers
- [ESG4PMChange+](https://esg4pmchange.com/) (Erasmus+, 2024 to 2027): ESG in project management education
- [HealthChain](https://healthchain-i3.eu/) (I3, 2023 to 2025): Leading SME for the FallPredict challenge with KBC Rijeka (StabilityGuard fall prediction)
- [Open datasets](https://github.com/SparkyScience/sparky-open-datasets) (CC BY 4.0): ESG LinkedIn Jobs Dataset; HealthChain Real-Time Health Sensor Dataset
- Academic partners: University of Split, University of Zagreb, University of Edinburgh, FKPV Celje, KBC Rijeka, SEA-EU

## Optional

- [Full company profile for AI agents](https://sparky.science/llms-full.txt): every case study with client type, approach and result
- [Open positions](https://sparky.science/open-positions/)
- [GitHub](https://github.com/SparkyScience), [LinkedIn](https://linkedin.com/company/sparkydata), [YouTube](https://www.youtube.com/@SparkyScience)