Packages
Productised engagements, transparent pricing
Every price is a 'from €X' starting point and value-based — the exact number falls out of a short scoping call once the work is clear. No rigid rate card, no surprises.
Architecture / AI-Readiness Audit
~2 weeks
A senior-led, fixed-fee read on whether your AI plan survives production — and what to fix first.
- Architecture & infrastructure review
- AI gap analysis + cost/latency reality-check
- Prioritised roadmap with effort/impact
- Written report + walkthrough call
PoC / Discovery Sprint
2–4 weeks
A working proof-of-concept against your real data, with an honest read on feasibility and cost at scale.
- Scoped, deployable PoC against your data
- Eval set + measured quality/latency/cost
- Go / no-go recommendation with numbers
- Path-to-production plan
RAG / LLM Build
4–8 weeks
A production retrieval system grounded in your corpus — evaluated, guarded, and fast enough for your UX (incl. Hebrew/multilingual).
- pgvector schema + chunking/retrieval policy
- Hybrid search + re-ranking + citations
- Self-hosted embeddings option (sub-50ms)
- Eval set, guardrails, latency/cost dashboard
Computer Vision / Multimodal Build
4–8 weeks
A production vision or multimodal pipeline — accurate, observable, and cost-gated so it's affordable to run continuously.
- Real-time video/RTSP or image pipeline
- Vision-model selection + motion-gating for cost (70–90% savings)
- Evidence extraction & alerting, or generation output
- Edge / on-prem or cloud deployment
Agentic Build
4–10 weeks
A tool-calling agent with the orchestration approach benchmarked for your use case — and an eval harness that proves it works.
- Orchestration + tool-calling design
- Framework choice benchmarked vs Pipecat/LiveKit/Vapi/Retell
- Task-completion eval harness
- Cost/latency observability
Claude Code Enablement
½–3 days
Get your team productive with Claude Code — and give your security team the Bedrock/VPC governance story to approve it.
- Hands-on workshop, custom to your codebase
- Bedrock + VPC deployment pattern
- Governance: review gates, repo risk tiers (RACI)
- Rollout playbook + ROI measurement
Fractional AI Architect
3–6 month minimum
A standing senior AI seat — architecture, reviews, and direction — at one to three days a week.
- ~1 day/week from €3,500/mo; 2–3 days from €7–10k/mo
- Architecture ownership + design reviews
- Vendor/model decisions + cost control
- Excludes token/tooling spend
Ad-hoc advisory: from €175/hour or €1,400/day for review, architecture and troubleshooting. For US-headquartered clients I bill in USD. All prices exclude VAT and travel.
Tell me what you're building
Book a 15-minute call. I'll tell you which engagement fits — or that it doesn't, if it doesn't.