🏢 Company facts
- Name: Synlogica
- Product: Synlogica Terminus
- Decision engine: GRIP (proprietary)
- Modules: Quality · Transport · Negotiation
- Founded: 2025
- HQ: European Union
- Hosting: EU default (AWS Frankfurt / Hetzner Falkenstein) · US optional
- Stage: Paid pilots
- Industry: Pharmaceutical decisioning software
- Compliance: GAMP 5 Cat 4 · GDPR · 21 CFR Part 11 compatible · EU GMP Annex 11
- LinkedIn: linkedin.com/company/synlogica ↗
- Email: info@synlogica.ai (general) · adam.karpinski@synlogica.ai (CEO)
👤 Founder bio
Adam Karpiński — Founder & CEO, Synlogica.
Years inside pharmaceutical supply chains and procurement. Started Synlogica after seeing the same pattern repeat — operational teams sitting through multi-day excursion reviews, tender leakage nobody could pin down, negotiation arguments rebuilt from scratch each time. Synlogica Terminus is the platform he wished existed: governed decisioning with a reproducible audit trail.
🎯 30-second pitch
Pharma decisions that should take minutes still take days — temperature excursions reviewed via Excel and email, tender leakage spotted only when invoices arrive, supplier price claims countered with gut-feel arguments. Synlogica Terminus brings quality, transport and supplier negotiation onto one governed platform. Decisions ship as audit-ready packages — inputs, model version, rationale, approvals, all SHA-hashed. Transport and Negotiation run on GRIP, Synlogica's proprietary probabilistic engine (Monte Carlo + Bayesian + Arrhenius). EU-default hosting, US optional. Currently running paid pilots.
🎨 Logo & brand
Brand wordmark in two contexts. The hex-"S" monogram favicon sits on dark — legible on both light and dark backgrounds.
🎨 Brand colors
#1f5fa6
#9bc2e6
#163f6e
#1b1d21
Clinical Blue is the primary brand colour; Soft Blue and Deep Blue support it, Ink anchors text and dark bands. Status colours (RELEASE green · CONDITIONAL amber · HOLD red) are constant in both themes.
🔬 Key technical facts (no jargon required)
- Probabilistic, not deterministic: GRIP returns distributions of outcomes with confidence intervals — not point estimates. Auditors get the math, not "the model said so".
- Three pillars: Hard rules (your SOPs/GxP) + probabilistic models (Monte Carlo / Bayesian / Arrhenius) + decision policy layer (your risk tolerance).
- Reproducible: Every decision package logs engine version, parameter version, rule version, policy version + SHA-256 hash. Replay = same output, bit-for-bit.
- NOT an LLM: No large-language-model in the decision loop. LLMs only parse free text into structured fields; their output never enters the recommendation.