SERVICES

Nalu AI. And more.

Alongside Nalu AI, individual solutions are built for data integration, machine learning and backend development — tailored, production-ready, maintainable.

MAIN PRODUCT

Nalu AI

Demand Intelligence Platform

ML forecasting, supply chain and sales intelligence — end-to-end from forecast to production. On-premise, config-driven, tailored for mid-market companies.

→ nalu-ai.com

customers in system

SKUs forecasted

time series processed

weeks to go-live

REFERENCE PROJECT

MeatMind

Built for a mid-market meat processor — in production every day. From demand forecasting (top-down ML, 8.1% WMAPE) through purchasing and slaughter planning to a supply-chain control tower, on top of SAP data.

Top-down ML forecastPurchasing & productionControl towerAnomaly detection

Python · LightGBM · PyTorch · FastAPI · DuckDB · Dagster · Docker · SAP

"The applications developed have become a permanent part of our daily operations."

CEO

Food producer · DACH mid-market

app.nalu-ai.com/kunden

KUNDEN-INTELLIGENZ

Wer kauft wann — und was.

Kaufbereit (30T)

143

Erw. Umsatz

€184k

Churn-Risiko

KundeKaufwahrsch.TerminStatus

REWE Group

KD-001 · Champion

92%

KW 20

Kaufbereit

Edeka Südbayern

KD-002 · Loyal

78%

KW 21

Kaufbereit

Gastro Service Nord

KD-047 · At Risk

34%

—

Churn-Risiko

Metzgerei Huber

KD-089 · Loyal

61%

KW 22

Bald kont.

770 Kunden · aktualisiert täglich 06:00 · Modell: Survival Analysis

PLANNING & CONTROL

The core: from forecast to the production line.

app.nalu-ai.com/steuerung

STEUERUNGSTOOL · S&OP-COCKPIT

Von der Prognose bis zur Produktion

142 t

Bedarf

Absatz-Forecast

→

Einkauf

Bestellvorschläge

→

86 %

Produktion

Kapazität

→

Control Tower

Engpässe

Kapazitäts-Auslastung · Linien

okengÜberlast

Zerlegung7280689160

Wolf8896847870

Verpackung6474827058

Kutter9098768864

MoDiMiDoFr

⚠ Bedarf 142 t · Eigenproduktion 128 t · Engpass KW 30 — Zukauf 38 empfohlen

Bedarf − Eigenproduktion − Zukauf = Restbedarfrollierend · 21 Tage

MeatMind doesn't stop at the forecast. The prediction is traced back into purchasing and production — a continuous control tool instead of isolated numbers. Demand, in-house production and external buy-in come together in an S&OP cockpit, with capacity signals and bottleneck early warnings.

Demand → purchasing (MRP) → production → control tower
Slaughter & line planning from the sales forecast
Purchasing matrix: demand − in-house − buy-in = net requirement
Capacity utilisation & bottleneck signals, rolling

INDUSTRIES

One product, several industries.

The platform is industry-agnostic at its core — each industry is a configuration. Proven in meat processing, ready for bakeries and beverage retail.

app.nalu-ai.com/forecast

PRODUKTGRUPPE · RIND

Rinderhack 5 % · Frischtheke

PG1-42 · 12-Wochen-Forecast · Top-Down

Modell · LightGBM

WMAPE · 8,1 %

Genauigkeit · 91,9 %

Absatz-Forecast · KW 24 – KW 32

Historie Forecast

142

Schlachtbedarf

Rinder/Wo

+38

Einkauf (Rest)

Rinder

Control Tower

Engpässe

Produktionsboard · Linien-Auslastung

Zerlegung

82%

Verpackung

68%

Kutter

96%

⚡ Hinweis: Grillsaison-Peak KW 29 · Schlachtplan angepasst

Standorte: Werk A · Werk BRetraining: Fr 03:00 Uhr

REFERENCE · IN PRODUCTION

Meat processing

MeatMind · live in production

The proven reference: demand forecasting at product-group level (top-down ML), traced back down to slaughter and purchasing needs.

Top-down forecast at 8.1% WMAPE, 12-week horizon
Slaughter & production planning from the sales forecast
Purchasing matrix with supplier ranking
Anomaly detection & local AI explanations (on-premise)

READY TO DEPLOY

Bakery

Daily freshness instead of overproduction

Daily per-store forecast with just one day of lead time — for bakeries that bake fresh each morning and want to minimise returns.

Daily forecast, 14-day horizon, 1-day lead time
Store heterogeneity (city centre, station, residential)
Weekend, holiday & seasonal patterns (Stollen, pretzels)
Production via recipes (BOM) & oven capacity

app.nalu-ai.com/filialen

FILIALE · TÄGLICHE PROGNOSE

Backwaren-Bedarf je Tag

InnenstadtBahnhofWohngebiet

Horizont · 14 Tage

Vorlauf · 1 Tag

Frische · same-day

Absatz je Wochentag

Wochenend-Peak

⚠ Mehl Type 550 · Reichweite 1 Tag · Bestellvorschlag offen

Rezept-Explosion (BOM) aktivOfen-Auslastung 86 %

app.nalu-ai.com/analytics

SAISONALITÄT · ABSATZ ÜBER DAS JAHR

Sommer · Oktoberfest · Q4

BierLimoWeinSektWasser

Artikel · 94 SKUs

Kategorien · 9

Horizont · 4 Wochen

Absatzindex · 12 Monate

Peaks: Sep · Dez

~100

B2B-Kunden

Churn-Risiko

96 %

Lieferanten-OTIF

Lifecycle: 3 Neu · 1 AuslaufAktualisiert täglich 06:00

READY TO DEPLOY

Beverage retail

Seasonal demand & B2B customers under control

Weekly forecast across 94 SKUs and nine categories — including B2B customer intelligence and pronounced seasonality.

Weekly forecast, 4-week horizon, 94 SKUs / 9 categories
Seasonality: summer, Oktoberfest, Q4 (sparkling & wine)
B2B customer intelligence: RFM, churn risk, revenue forecast
Supplier scorecard (OTIF, MOQ, lead time)

Bakery and beverage retail are ready-to-deploy industry configurations with realistic sample data. The figures shown in the views are illustrative.

OTHER SERVICES

What else gets built.

Project-based. On request.

ETL & Data integration

Pipelines from SAP, SQL Server, heterogeneous source systems. Automated, production-ready, maintainable.

Custom ML

Forecasting, classification, anomaly detection — plus LLM/AI integration, either local (on-premise) or cloud, for chat, reports and automatic insights.

Backend & APIs

FastAPI, Flask, REST interfaces. From concept to deployment on your own infrastructure.

DWH & Data architecture

DuckDB, PostgreSQL, Parquet pipelines. For companies without a central data foundation or with grown structures.

Data Analytics & Reporting

Dashboards, KPI analysis and ad-hoc queries from SAP, SQL and ERP systems — as an interactive dashboard or automated report.

Operational Automation

Scripts for recurring processes: data exports, report generation, Windows automation, scheduled tasks.

Pricing on request — every project is different.

app.nalu-ai.com/pipeline

DATENPIPELINE · NÄCHTLICH

Von der Quelle bis zur Analyse

SAP R/3SQL ServerCSV / Excel

→

Sync

Rohdaten → Parquet

→

Transform

bereinigt · validiert

→

Load

DuckDB

Letzter Lauf · heute

02:00✓sync · SAP R/31,24 M

02:18✓transform · Regeln1,24 M

02:31✓load · DuckDB1,24 M

02:33✓quality-check3 Lücken gefüllt

Read-only · keine Änderung am QuellsystemDagster · orchestriert

LIVE VIEW

ETL & Data integration

Pipelines from SAP, SQL Server, heterogeneous source systems. Automated, production-ready, maintainable.

LIVE VIEW

Data Analytics & Reporting

Dashboards, KPI analysis and ad-hoc queries from SAP, SQL and ERP systems — as an interactive dashboard or automated report.

app.nalu-ai.com/dashboard

Übersicht

KW 18 · Mo, 8. Mai

7T30TYTD

Gesamtabsatz

€ 2.4 M

+8.2 %

Forecast-Genauigkeit

96.2 %

MAPE 7.6 %

Service Level

98.1 %

+1.4 %

Aktive Forecasts

894

12 neue

Absatz vs. Forecast · 12 Wochen

LightGBM · MAPE 7.6 %

Historie Forecast

ABC/XYZ-Matrix · 894 Artikel

automatisch klassifiziert

142

134

112

198

A · Hoher Umsatzanteil

B · Mittelwert

C · Geringer Umsatz

XYZ = Variabilität · niedrig → hoch

app.nalu-ai.com/jobs

AUTOMATISIERUNG · SCHEDULER

Läuft, ohne dass jemand dran denkt

5 aktiv

JobZeitplanNächsterStatus

Wochenbericht → E-Mail

Ø 4 s

Mo 06:00Mo 06:00✓

Forecast-Export → SAP

Ø 11 s

tägl. 05:0005:00✓

Reorder-Alerts → Teams

Ø 2 s

tägl. 07:3007:30✓

Stammdaten-Abgleich

Ø 38 s

stündl.14:00!

Backup → Off-Site

Ø 1 min

tägl. 23:0023:00✓

Retry · Alert bei FehlerWindows Task · Cron · Airflow

LIVE VIEW

Operational Automation

Scripts for recurring processes: data exports, report generation, Windows automation, scheduled tasks.

Request a project.

No form. No funnel. Just an email.

aloha@nalu-ai.com

Or LinkedIn: /in/maximilian-fischer-naluai

Pricing on request — every project is different.