builder of systems

Danial Jamali

20 · industrial engineering

I design AI products, automation infrastructure, and analytics platforms — built to run in production, not in demos.

AI Systems · Data Intelligence · Automation Infrastructure
running in production

MathematicsIndustrial EngineeringPythonData AnalysisData MiningData EngineeringMachine LearningBusiness IntelligenceAI Products
01 / 09 · mathematics

moment 03 · hooshmand

Fly inside a single reply — understand, retrieve, reason, compose, respond.

hooshmand · AI culinary intelligence system · productionv3.0
skills
  • intent.parserfa/en
  • recipe.retrieval9k
  • pantry.matchon
  • nutrition.guardon
  • dialog.managerv3
turn #2241 · streaming5 stages
HM-001understand · request & dietInterpreting user intent and dietary constraints.96 ms
HM-002retrieve · recipes & pantrySearching recipes and pantry context.284 ms
HM-003reason · plan & substituteBuilding a feasible cooking plan.511 ms
HM-004compose · steps & timingGenerating steps and timing.173 ms
HM-005respond · stream replyStreaming the final answer.88 ms
telemetry
pipeline
standby
latency
—
retrieval
—
reasoning
—
model
loaded
stream
idle
version
v3.0
inference trace · live
warming the model

moment 04 · hooshmand · live

A head chef in chat — suggests food from what you have and what you want.

Hooshmand (هوشمند) is a live culinary intelligence system that acts as head chef — people open it, ask what to cook, and get dish ideas tailored to their ingredients, time, and diet.
Built on retrieval, reasoning, and real-time ingredient matching.

hooshmand · culinary intelligence system · production● production
Hooshmandهوشمند · head chef · v3
role
head chef
recipes
indexed recipes
pantry
match ingredients
languages
fa / en

moment 05 · sales dashboard

The interface builds itself.

A business intelligence and analytics platform built to automate reporting, forecasting, monitoring, and decision-making.

business intelligence platformproduction · live
1Mevents processed
47active automations
18live dashboards
99%uptime
revenue · 12 months+18% YoY
workflowstatusrecords
Inventory Synchourly · ETLlive284,000
Demand Forecastml pipelinerunning47
Revenue Reportauto · 12m agofresh18
Customer Segmentationnightly batchscheduled92,400

moment 06 · business intelligence

Raw data becomes decisions.

Centralizing fragmented operational data into a single decision-making layer — forecasting, reporting, and executive visibility.

decision intelligence · executive workspaceproduction · live
14data sources
26active reports
12automated pipelines
78%decision coverage
analytics output · forecast · reports · adoption6 months
data source mixconnected
  • ERP · 34%
  • CRM · 26%
  • Sheets · 22%
  • APIs · 18%
decision coverage · by domain5 functions
Finance92
Operations71
Forecasting83
Reporting64
Executive88
  • decision register · executive briefings
  • exec · board KPI summary12m ago
  • forecast · demand planning41m ago
  • finance · P&L rollupscheduled
  • ops · inventory alert1h ago

moment 07 · data

From raw data to production decisions.

Messy data in — forecasts, reporting, and decision tools out. Production pipelines, not one-off notebooks.

pipeline · fact_orders · warehousepython · sql · pandas · sklearn
In [1]: df = pd.read_sql("SELECT dt, amount, qty FROM fact_orders WHERE dt >= %s", engine, params=[since])In [2]: daily = df.groupby("dt").agg(revenue=("amount", "sum"), units=("qty", "sum"))In [3]: feat = engineer_features(daily, lags=(7, 28)) # ETL · feature storeIn [4]: mape, fcst = backtest_forecast(feat, horizon=14, model="ridge")MAPE 8.2% · 14-day forecast · published to reporting
driver analysis · spend vs revenueMAPE 8.2%
feature engineeringforecastingETLproduction pipelinesbusiness metrics

moment 08 · automation

Agents reason — then act.

Autonomous execution across systems — trigger, decide, act, deliver outcomes. Not generated text; production actions.

agent execution · order-ops · production● production
6connected systems
8active workflows
4trigger sources
3execution pipelines
Webhooknew order
Codepython · enrich
HTTPCRM lookup
AI Agentclassify · decide
n8nroute · branch
Slacknotify team
reasoning layer · pythonlive
from agents import classify, act def handle_order(event: dict) -> dict:    ctx = enrich(event)    decision = classify(        ctx, model="decision-model",        tools=[crm, inventory]    )    return act(decision, systems=["erp", "slack"])
trigger sourcesagent reasoningsystem actionsexecution pipelines
AI Product
Business Intelligence
BI Platform
Data Engineering
Automation Systems
Knowledge

build engine

One builder. Many systems.

Every project is an output of the same operating system — observe, analyze, decide, execute.

Danial Jamali

The technology fades. The builder remains.

I build systems. From large-scale data analysis and business intelligence to automation pipelines, production Python, and AI-powered products. The goal is always the same: turn complexity into decisions, and decisions into execution.

moment 11 · contact

If the problem is hard, I want to hear it.

AI systems, automation, analytics, architecture — or how any of this was built. Write below. I read every message.

secure channeldirect to me