>_ Tibor Sloboda

AI consultant & research engineer · slovakia

Tibor SlobodaAI research · systems engineering

I help people work out where AI is useful, build the systems behind it, and solve the technical problems that get in the way. Models, data, software, hardware, and the business they need to serve.

fig. 0 · loss surface ℒ(θ); gradient descent w/ momentum, 263 steps. converged.

My work spans AI research, production software, and the hardware underneath. I have led AI engineers, co-developed production data and LLM systems, and worked on research in medicine and chemistry. I care whether an idea is useful, whether it runs under your constraints, and whether it makes economic sense.

§1Start with your problem

See pricing →
01 We know what we need. We need someone to build it.You have clear requirements, a specification, or a roadmap. I can implement the work, lead your engineering team, or do both. 02 Would AI actually help here?An idea, some data, or a business problem. I assess what is feasible, what it would cost, and whether AI is the right tool at all. 03 It works, but it needs to work better.Slow inference, unreliable models, difficult data, cloud bills, or a system that keeps breaking between layers. I investigate the cause and implement the fix. 04 We need something unusual built or investigated.A research result to implement, an unfamiliar domain, an unusual model or hardware constraint. This is work I enjoy. 05 Our team needs to understand and use this.Practical training shaped around your team and its work, including AI fundamentals, LLMs for software engineering, and safe use of AI tools.

Flexible rates, excluding applicable VAT. Free 15-minute introductory call.

§2The corpus

fig. 1 · computed from this site's text
Corpus map of this siteEvery frequent term from the site text, embedded in 2-d from PPMI co-occurrence statistics and grouped into 5 clusters. 5 capability terms are highlighted as links.workresearchsystemsspandatamodelclassdivproductionfiguresoftwarehardwareindexmodelssystemproblemwhereslovakservicesdeploymentengineeringleadershipevaluationfigcaptionnbspfocus-liststrongtechnicalincludesstartratesboundarylogsdata-areawhetherneedshelplearninginfrastructurejunepolicyactualoperatingthresholdmdashmedicinechemistryhowincludingcalldiscoverymachineimplementationselectedinformationuniversitytechnologyfocusfiitdifferentresultstoolslatencyarchitecturetoolllmworkedconstraintsteamsbuildingpredictionpeoplewhytelltimeconsultingphdruntimeintendedessayspathhourspossibleslovakiacontactengineeravailablepresentLocal InferenceAI GuardrailsRust InfrastructureBiomedical MLProduction MLOps
fig. 1 · a vocabulary map built from the site’s text using PPMI co-occurrence, spectral projection, and five clusters. Ringed nodes link to technical topics. Positions are adjusted for legibility.

§3Areas of expertise

fig. 2 · experience profile

My depth of experience across six areas of research and engineering.

Areas of expertiseSelf-assessed depth of experience. Machine learning: 9/10; Data engineering: 8/10; AI infrastructure: 6.5/10; Software engineering: 8/10; Security: 6.5/10; Multidisciplinary research: 10/10.MachinelearningDataengineeringAIinfrastructureSoftwareengineeringSecurityMultidisciplinaryresearch0246810986.586.510
fig. 2 · self-assessed depth of experience, on a scale from 0 to 10.

§4Selected work

full portfolio →

I led the team and co-developed an automated pipeline turning heavily unstructured, inconsistent inputs into standardized business data.

I led the team building a multi-site, multi-agent LLM chatbot with continuous memory and presence, taking responsibility for technical decisions and prototypes.

I built a GPU cluster for an AI engineering team's experiments and managed and administered the hardware.

§5Research & writing

Published research, software experiments, and opinions I have spent too long thinking about.