TELCOTEC

Software Engineering · Data Platforms · Applied AI · UK consultancy since 2004

Services

What we do, and how we do it.

Five practices, with technical advisory across all of them. For each practice: when to call us, what we do and what you get.

Practice 01

Software & SaaS Engineering

We design and build web applications, customer portals and SaaS platforms, and take over existing systems that have become hard to change. The same senior people stay with the work from first design to production.

Where we have done this

eMock, an online exam preparation platform, was designed and built by Telcotec: accounts, subscriptions, automated marking and AI-assisted learning.

About eMock

When to call us

  • You have an idea for a new product or portal and need it built properly the first time.
  • Your existing application has become slow, fragile or expensive to change.
  • You want to turn an internal tool into a product other organisations can subscribe to.
  • You need disconnected systems to work together.

What we do

  1. Discovery and architectureWe agree the users, the must-haves and the constraints, then design an architecture that fits your size and budget.
  2. Short build cyclesYou see working software every week or two, not a big reveal at the end.
  3. SaaS foundationsSecure sign-in, user roles, multi-tenant data, subscriptions and payments, built in from the start.
  4. Integrations and APIsClean, documented APIs and connections to the systems you already use.
  5. Testing and releaseAutomated tests and repeatable deployments, so changes are safe to make.

What you get

  • A working application in production
  • Source code in your repository, with documentation
  • Automated tests and a deployment pipeline
  • A handover so your team, or ours, can run and extend it

Technologies we often use

  • .NET / C#
  • PHP / Laravel
  • Python
  • React
  • TypeScript
  • REST APIs
  • Multi-tenant SaaS

Practice 02

Data Engineering & Analytics

We bring data from many systems into one reliable place, then build the reporting on top, so the numbers agree and answers come back in seconds.

Where we have done this

For telecom operators we built NPP, which brings call records from multi-vendor switches into one view and turns them into near-real-time reporting at volumes of 150 million records a day.

About NPP

When to call us

  • Data sits in several systems and nobody trusts the totals.
  • Reports take minutes, or run overnight, when people need answers now.
  • Spreadsheets have quietly become business-critical software.
  • Data volumes are growing faster than your current setup can handle.

What we do

  1. Data pipelinesWe collect, clean and load data from your sources, on a schedule or in near real time.
  2. One data modelAgreed definitions, so every report counts the same thing the same way.
  3. Fast analytics storesFor high volumes, columnar databases such as ClickHouse that are built for fast aggregation.
  4. Reports and dashboardsStandard reports, dashboards and ad hoc analysis that people actually use.
  5. Data quality checksAutomatic checks that catch missing or duplicated data before it reaches a report.

What you get

  • Documented pipelines that run without manual effort
  • A clear data model with agreed definitions
  • Dashboards and reports your team can maintain
  • Alerts when data is late or wrong

Technologies we often use

  • ClickHouse
  • SQL Server
  • PostgreSQL
  • Python
  • ETL / ELT
  • Streaming design
  • Dashboards

Practice 03

Database Performance & Modernisation

We find out why a database is slow or expensive, fix what can be fixed quickly, and give you a measured plan for anything bigger: tune, re-architect or migrate. We work with open-source and commercial databases, and recommend whichever fits your data, team and budget.

Where we have done this

In a benchmark study on production-scale test data at telecom-operator volumes, we tested whether ClickHouse could take over a SQL Server reporting layer without changing a single result.

Read the case study

When to call us

  • Queries time out, or reports slow down as data grows.
  • Storage and licence costs keep rising.
  • You are considering a new database platform but want evidence before committing.
  • The platform has hit its limits and patches no longer help.

What we do

  1. Performance reviewWe measure where time and money actually go: slow queries, indexes, storage, hardware and licences.
  2. Quick winsQuery tuning, indexing, partitioning and columnstore changes that can make a difference within days.
  3. Open source or SQL ServerOpen-source databases such as PostgreSQL or ClickHouse have no per-core licence fees. Sometimes staying on SQL Server and tuning it is the better answer. We show you the trade-offs.
  4. Evidence before migrationWe run old and new side by side on your own data and check that every result matches, value by value.
  5. A low-risk migration planStep by step, with checkpoints and a way back at each stage.

What you get

  • A written findings report with measured results
  • Quick fixes applied and verified
  • Costed options: tune, re-architect or migrate
  • A migration plan, if you choose to move

Technologies we often use

  • SQL Server
  • PostgreSQL
  • MySQL / MariaDB
  • ClickHouse
  • Columnstore
  • Partitioning
  • Query tuning

Practice 04

Cloud Architecture

We design, build and review cloud environments that are secure, scale with demand and cost what you expect them to. We work on the platform you already use, whether that is Microsoft Azure, Amazon Web Services or Google Cloud, and where cloud is not the right fit, we design for your own servers instead.

Where we have done this

NPP is installed on operators' own infrastructure, and eMock runs on Azure. We design for both, and for the day-to-day running of a system, not just its launch.

Our products

When to call us

  • You are moving applications to the cloud and want it done properly.
  • Your application needs to handle more users or more data.
  • Your cloud bill has grown and nobody is sure why.
  • Deployments are manual and risky.

What we do

  1. Architecture designThe right mix of services for your workload: app hosting, serverless functions, containers and managed databases, on Azure, AWS or Google Cloud.
  2. Security and accessIdentity, networking and secrets set up so only the right people and systems get in.
  3. Automated deploymentsCI/CD pipelines and infrastructure as code, so every environment is repeatable.
  4. Cost reviewWe find waste, right-size resources and set up budgets and alerts.
  5. MonitoringLogs, metrics and alerts, so problems are found before users notice.

What you get

  • An architecture document and diagram
  • Infrastructure defined as code in your repository
  • Automated build and release pipelines
  • A cost report with savings, and monitoring in place

Technologies we often use

  • Microsoft Azure
  • Amazon Web Services (AWS)
  • Google Cloud
  • Containers & Kubernetes
  • Serverless
  • CI/CD
  • Infrastructure as code
  • On-premises servers

Practice 05

Applied AI & Automation

We add AI where it helps your users or your team, grounded in your own checked data, with guardrails so it doesn't give wrong answers or leak sensitive information.

Where we have done this

Rocky, our AI assistant, is built into both NPP and eMock. In NPP it answers network performance questions from checked queries with sensitive data masked. In eMock it helps students with vocabulary, comprehension and essays, and explains answers.

Our products

When to call us

  • You want customers or staff to ask questions in plain English and get reliable answers.
  • Your team spends hours on repetitive work such as marking, summarising or data entry.
  • You have tried AI tools but can't trust the answers.
  • You want to know what AI could do for you before you spend money.

What we do

  1. The right use caseWe pick one problem where AI clearly adds value and agree how success will be measured.
  2. Grounded in your dataAnswers come from your checked data and approved queries, not guesses.
  3. Guardrails and privacySensitive data masked, answers checked, and questions and responses logged for audit.
  4. Prototype firstA working prototype on real data, so you can judge it before committing to a full build.
  5. Built into your workflowAI features inside your existing application and processes, not a separate tool.

What you get

  • A working prototype tested against agreed acceptance criteria
  • Guardrails, data masking and audit logging
  • Measured results: accuracy, time saved or cost
  • A plan to take it to production

Technologies we often use

  • Large language models
  • Retrieval (RAG)
  • Ask-your-data
  • Data masking
  • Audit logging
  • Python
  • Workflow automation

Across all five

Technical Advisory

Independent, senior advice for decisions you want to get right before you commit budget.

Get in touch →
  • Feasibility studiesCan it be done, how, and roughly what will it take?
  • Architecture and code reviewsA clear view of strengths, risks and what to fix first.
  • Technical due diligenceFor investors and buyers who need to understand a system before a deal.
  • RoadmapsA practical sequence of steps, with priorities and rough effort.

Technologies

The right tools for your situation.

We are not tied to one vendor. We work with open-source and commercial technologies, in the cloud or on your own servers, and recommend what fits your data, your team and your budget.

Applications

Web applications, portals, SaaS platforms and APIs.

  • .NET / C#
  • PHP / Laravel
  • Python
  • React
  • TypeScript
  • REST APIs

Databases

Open source or commercial: we recommend what fits your data, team and budget.

  • SQL Server
  • PostgreSQL
  • MySQL / MariaDB
  • ClickHouse

Data & analytics

Pipelines, analytics stores and reporting.

  • ETL / ELT
  • Streaming design
  • Columnar analytics
  • Dashboards & reports

Cloud & hosting

On the cloud platform you already use, on your own servers, or a mix of both.

  • Microsoft Azure
  • AWS
  • Google Cloud
  • Containers & Kubernetes
  • CI/CD
  • Infrastructure as code
  • On-premises

AI

Practical AI with guardrails, grounded in your data.

  • Large language models
  • Retrieval (RAG)
  • Data masking
  • Audit logging

How we work

Clear from the first conversation.

  1. 1A short callWe listen to your goals, constraints and data.
  2. 2A written scopeWhat we will deliver, how we will measure it and what it costs, agreed before work starts.
  3. 3Build in incrementsWorking results every week or two, with room to change course.
  4. 4Handover and supportDocumentation, a walkthrough, and ongoing support if you want it.

The exact deliverables for each project, such as monitoring, infrastructure as code and handover, are agreed in the written scope before work starts.

Anything we develop for you is yours. If we build on our own existing tools, such as NPP, you get a licence to use them.

Work with us

Have a project in mind?

Tell us what you need. We will reply within one working day and suggest a sensible first step.

Get in touch →