About DataWise

Why DataWise exists#

There is a gap in this market, and most companies fall into it. On one side, AI that demos beautifully and never survives contact with your ERP, your data, or your Tuesday morning. On the other, transformation programs sized for organizations with a thousand employees and a steering committee. In between sit most Israeli manufacturers, importers, distributors, and operations companies: real problems, real data, and nobody offering to actually build the system and run it in production.

DataWise exists for that gap. We are a boutique consultancy, deliberately small, and we build production systems, not demonstrations. The demo is the easy part. The finish line is a system running inside your environment, integrated with the tools you already run, measured against a success metric we agreed on before anything was built.

One structural fact matters more than any adjective on this page: at DataWise, the person who characterizes your problem is the same person who builds the solution and integrates it into your systems. Industry reporting keeps finding the same stall point in AI projects, the handover, where the people who scoped the work pass it to people who were not in the room. We removed that gap by not having a handover at all.

What “multi-disciplinary” actually means#

Every consultancy claims breadth. Here is what it means concretely here: the same engineer who builds a forecasting ensemble for your demand planning also architects the closed-network deployment it runs in. The same person who trains a vision model to segment aerial imagery builds the correction UI your analysts will use, and the pipeline that keeps it fed.

The disciplines under one roof:

• Time-series forecasting: demand, sales, commodity prices, cashflow, freight. Ensembles of statistical, gradient-boosted, and deep models, graded openly against actuals.
• Computer vision: segmentation, measurement, and analysis of imagery, wrapped in tooling humans can supervise and correct.
• Online and reinforcement learning: systems that keep learning after delivery instead of decaying quietly.
• Data science and analytics: churn analysis, anomaly detection, customer clustering, explainability that leadership can act on.
• Production systems engineering: cloud-native platforms, APIs, MLOps, integration with the systems you already run.
• Air-gapped and cloud deployment: the same capabilities, whether your constraint is a fully isolated closed network or a modern cloud stack.

Why this matters to you: an AI project is never just a model. It is data plumbing, integration, deployment, monitoring, and a user who has to trust the output. When one senior engineer holds all of those layers, nothing gets lost between specialists, and nobody tells you “that part is someone else’s problem.”

Honesty is the operating system, not a value statement#

Three practices define how we work, and you can verify each one:

• Graded ledgers. A forecast nobody checks afterwards is an opinion. Wherever our systems make predictions, we track them openly against actuals, month after month. On our longest engagement, a three-year forecasting service for a US-market industrial manufacturer, every monthly forecast was graded against real results with 3-month and 12-month error tracking, and that ledger sat on the table in leadership planning meetings.
• Calibrated verbs. Every piece of work we describe on this site uses the highest truthful verb and nothing above it: “built, delivered, and ran for years” only where that happened; “built and validated on real data” where that is the truth; “designed and architected” or “consulted and advised” where that is. If we only designed something, we will not tell you we shipped it.
• “Not yet” is a legitimate answer. Characterization can conclude that AI is not the right tool for your problem, or that your data is not ready. When it does, we say so in writing, with what to fix first. We would rather lose an engagement than deliver a system that quietly fails you.

The founder#

Dan Malka is the founder and principal of DataWise. His signature line on scope-of-work documents reads, in his own words: יועץ, מומחה לשילוב מדע-נתונים עסקי (consultant, specialist in integrating data science into business).

Dan personally leads every engagement: the characterization, the build, and the delivery. His work spans multi-year production engagements rather than one-off deliverables. Two anchor the record:

• A three-year embedded forecasting service for a US-market industrial manufacturer (Mueller-group), delivering monthly into leadership planning meetings, with every forecast graded against actuals; CEO recommendation on file.
• A full operations transformation, live and in daily production use: a live-performance company’s entire operation, previously scattered across WhatsApp groups and a dozen tools, centralized into one work-OS with dashboards, automations, and failure monitoring.

Beyond those, Dan has built and validated systems on real data across forecasting, vision, behavioral prediction, and LLM agents, and has designed and specified architectures for organizations from importers and distributors to security-sensitive closed-network environments.

What we are not#

We are not a staff-augmentation shop: if you need twenty developers, we are the wrong call and will say so. We are not an organization-wide transformation program: enterprise integrators do that, and do it well, for enterprises. We are not a chatbot vendor, and we do not resell a platform with a meter on it. What we build is yours: the models, the code, the pipelines. If that is not what you need, we will tell you in the first conversation, which costs you nothing.