Finance & banking
Risk, regulatory reporting and the operational data that sits underneath both.
Leamington Spa · United Kingdom

Independent data consultancy
We migrate, model and govern the systems your organisation already runs on - then help you see what they are telling you. Practitioner-led, based in Warwickshire.
2021
Founded in Warwickshire
20+
Years’ hands-on data craft
7
Service lines, one practice
G14
G-Cloud listed supplier

Cyber Essentials PLUS
G-Cloud 14 listed · Company no. 13791456
The practice
iDatamaze was founded in 2021 by Srinivas Lakkireddy after two decades in data engineering, Oracle estates and large migrations. We stayed small so clients speak to the people who write the SQL.
Integrity
We will tell you if the data cannot support the claim. That conversation is cheaper at the start.
Craft
The people you meet write the SQL, the models and the runbooks. Boutique on purpose.
Clarity
Plain language. Named owners. Definitions written down. If we cannot explain it, it is not finished.
Stewardship
Your data stays yours. We design for handover, not lock-in, and we treat access as a privilege.
Services
Cloud, models, security, insight, quality, applications and measurement - always in that order of honesty: data first, theatre never.
01
Move the estate. Keep the meaning.
Migrations, warehouses, lakes and pipelines - planned around the data you cannot afford to lose.
02
Practical models that stay in production.
AI models, machine learning and automation for UK organisations - built in Leamington Spa on clean data, with an owner, and a way to tell when they are wrong.
03
Protect the data, not just the perimeter.
Threat detection, identity, encryption and incident response - aimed at the data you are required to keep safe.
04
See the business as it actually is.
Catalogues, models, dashboards and the questions behind them - so leaders are not arguing over whose spreadsheet is right.
05
Trust the number before you act on it.
Tests for pipelines, models and applications - so a silent wrong figure does not become a public one.
06
The front door, connected to the books.
Sites and CRM that talk to your data - not another island with its own version of the customer.
07
Spend that you can follow to a result.
Campaigns and content with the measurement designed in - so you know what moved, not just what shipped.
How we work
We do not sell a methodology with a trademark. We do the next honest step, in a language you can take to a steering group.
We sit with the people who live the problem. Systems, constraints, politics, the question behind the brief. No workshop theatre.
Sources, owners, grain, quality, risk. A picture of the estate you can argue with - and then agree.
Pipelines, models, controls, the application if it is needed. In slices you can see, test and stop.
Runbooks, catalogues, named owners. We leave knowledge behind, not a retainer you cannot leave.

Sectors
We are not a vertical factory. We are useful where the data is heavy and the questions are real.
Risk, regulatory reporting and the operational data that sits underneath both.
G-Cloud 14 listed. Case data, performance, and services that have to stand up to scrutiny.
Patient, operational and research data - with security treated as a design constraint, not an add-on.
Network, usage and customer data at a volume that punishes a sloppy model.
Stock, demand and the customer - in one place, in time to act.
Product analytics, platform data and the plumbing growing companies outrun.
Student outcomes, funding and administration without a second set of books.
Supply chain, quality and plant data that has to match the shop floor.
Insights
Capacity bills, silent failures, near-live refresh, Power BI, Tableau, BigQuery, Oracle and GDS. Written the way we say it on a call.
Quality
The job succeeded. The number is still wrong. That is the expensive kind of failure.
Analytics
Every extra minute of freshness has a price. Most boards do not need ten seconds. They need a number that is still true at 09:00.
Analytics
The visual is the last step. Grain, relationships and a filter that nobody sees do more damage than a colour palette.
Cloud data
On-demand bytes, idle PX, and a dictionary nobody reads. Three ways to lose money while the dashboard stays green.
AI models
An agent is a loop with tools. If the model, the prompt and the data all move, you need a harness more than a demo.
Practice
We have done this for public and private clients. The pattern is the same: catalogue, grain, a path you can roll back, then a dashboard that uses the new meaning.
In their words
These are from people who hired the work, not a marketing panel. We have left them largely as they were written.
“Sri provided data analysis services and was an excellent contractor with a high technical ability. He has extensive experience in Oracle and large data-set migrations, solutions analysis and data catalogues.”
“I was impressed by Sri’s ability to deal with even the toughest projects effortlessly. He has great skill in handling data science work in Python. I strongly recommend him - a pleasure to work with.”
“Proactive and fully committed is what comes to mind. I was impressed by how quickly he understood the task and shared his expertise. A true asset.”
“I have always found Srinivas hard-working and dedicated. He works largely unsupervised through complex tasks. Nothing I have presented him with has seemed to daunt him.”
Questions
If yours is not here, write. We would rather a short honest email than a long form.
We are a UK data consultancy. The core of the work is migration, integration, engineering, analytics and the security around it. Cloud platforms, models and applications sit on that foundation - we do not start with a tool and hunt for a problem.
A short conversation is usually enough to know whether this is a two-week discovery or a longer piece of work. No pitch deck.