Diagnostic
We study one decision, the data behind it and how it is made today, then tell you plainly whether optimisation, ML or analytics would change the outcome.
Decision brief and a recommended path
From the first conversation to a system your team runs.
Each engagement centres on one decision: a schedule, a replenishment policy, a route plan, a budget split. We model it properly, prove it against your history, and put it where your planners already work.
- Scoped to one decision first
- Validated against your own history
- Built into the tools you use
Five stages, each with a deliverable.
You see working results early, and every stage ends with something your team can review before the next one starts.
-
Understand the decision
Who decides, how often, what "good" means, and which constraints are real.
Decision brief
-
Frame & model
Objectives, constraints, data and uncertainty, reviewed with the people who will use the answer.
Formulation & data map
-
Build & validate
Solve or train, then backtest against your current baseline and against weeks your team already knows well.
Validated model
-
Integrate & enable
Deploy into your planning cycle, systems of record or a purpose-built app, then train the team that uses it.
Running system
-
Improve continuously
Monitor decision quality and data drift, and retune when costs, capacity or demand change.
Monitoring & iteration
Three ways to start.
Pick the entry point that matches where you are. Most clients begin with a diagnostic and grow from there.
Build & integrate
We formulate, build and validate the model, then wire it into your planning cycle through an API, a scheduled job or a focused app for the planners.
A validated model in daily use
Run & improve
We monitor decision quality and data drift, retune when costs, capacity or demand change, and extend the model to the next decision.
Ongoing support and iteration
Proven tools, chosen per problem.
We work with established commercial and open-source solvers, ML frameworks and cloud platforms, and fit them to your existing stack.
Optimisation
- Gurobi
- CPLEX
- OR-Tools
- HiGHS
- Pyomo
- PuLP
ML & forecasting
- scikit-learn
- XGBoost
- LightGBM
- PyTorch
- StatsForecast
Simulation
- SimPy
- AnyLogic
- Monte Carlo
GenAI
- LLM APIs
- RAG
- LangChain
- Vector databases
Engineering
- Python
- SQL
- FastAPI
- Docker
- Airflow
- AWS · Azure · GCP
Questions we often hear.
Not covered here? Ask us directly.
What is operations research, in plain terms?
It is the science of making the best decision under constraints. You describe the goal (lower cost, higher service, less waiting) and the limits (capacity, time, rules), and a mathematical model finds the best plan among millions of options.
How is this different from BI or dashboards?
Dashboards describe what happened. Our work is prescriptive: it recommends what to do next, explains why, and re-plans when conditions change. It often sits on top of the BI and data you already have.
Do we need perfect data before we start?
No. Data cleaning and reconciliation are part of the engagement. We build models that tolerate real-world gaps, and we show which data improvements would matter most to the decision.
What does a first engagement usually look like?
We scope it to one decision, such as a schedule, a replenishment policy or a forecast. That way you see a working, validated model early, before anyone commits to a larger rollout.
Will it work with our existing systems?
Yes, that is the goal. We integrate with ERPs, planning tools, data warehouses and internal apps through APIs or scheduled jobs. Where a new interface helps, we build a focused web app for the planners.
Which industries do you work with?
Manufacturing, supply chain and logistics, retail and CPG, marketing and consumer brands, e-commerce, aviation, healthcare, pharma, energy, mining, financial services and service operations. Basically, any operation with real constraints and costly decisions.
What is Geniemix, and how does it relate to OR Genie?
Geniemix is our own product for marketing mix modelling, built by the OR Genie team and available at geniemix.in. It suits brands that want a weekly, self-serve answer on where their ad budget works. When a business needs more, such as offline media, retail and distribution effects, or a model inside its own systems, we deliver marketing optimisation as a consulting engagement.
Bring us the decision that is costing you most.
Tell us about the choice, the constraints and what a better answer would change. We'll reply with an honest view of whether optimisation, AI or analytics can help.