Practical AI and automation for real operational problems

I connect AI models, telephony, IT tooling and market data to the systems you already run, so work moves through automated workflows instead of manual hand-offs. This page covers what I build, how an engagement runs, and what you can expect.

Schedule a consultation LinkedIn
AI workflow automationLLMs, RAG, assistants, reporting Voice AI & call automationVoIP, PBX, transcription, analytics MSP & IT automationRMM, M365, AD, CRM, networking Algorithmic tradingMarket data, options chain, signals

AI workflow automation

Most of my work starts with an operational problem rather than a model. I connect LLMs, RAG pipelines and AI assistants to the systems a business already runs, so that reporting, data pipelines, customer support and internal operations move through automated workflows instead of manual hand-offs. The glue is usually Python: automation scripts and backend services that talk to REST APIs, databases and financial-data feeds.

What I build

  • LLM-powered applications for summarising, classifying, extracting and drafting.
  • RAG pipelines that answer from your own documents, tickets and knowledge base.
  • AI assistants for support and internal teams, wired into the tools they already use.
  • Automated reporting that pulls from several sources and delivers on a schedule.
  • Data pipelines that collect, clean and load data from APIs and databases.

Where it fits

  • Customer support triage and first-draft replies.
  • Internal operations: approvals, hand-offs, status updates.
  • Recurring reports that someone currently builds by hand.
  • Searching and answering over scattered documents.
  • Moving data between systems that have no native integration.

How I approach it

I start from the manual process: who does it, how often, what goes wrong. I then decide where an AI model actually helps and where plain code or a rule is more reliable. Where a model is involved, I keep a person in the loop for decisions that matter and build in logging, so you can see what the system did and why.

PythonLLM APIsRAGREST APIsDatabasesAutomation scriptsBackend services

Voice AI and call automation

With 20+ years in telecom and VoIP, I build voice assistants and call analytics that sit on top of VoIP and PBX systems. Calls can be transcribed, scored for sentiment and turned into call intelligence that teams can actually search and act on, without replacing the phone system they already have.

What I build

  • Voice assistants that handle routine inbound or outbound calls and hand off to people when needed.
  • Call transcription from your existing recordings or live call streams.
  • Sentiment and quality scoring across calls, not just a sampled few.
  • Searchable call intelligence: summaries, topics, follow-ups and action items.

Why the telecom background matters

  • I understand SIP, call flows, PBX configuration and recording setups, so the AI layer is built on how calls actually move.
  • It works alongside your current phone system rather than forcing a migration.
  • Call data can feed your CRM and ticketing tools directly.
Call recording and analysis involve privacy and consent rules that differ by country and industry. Part of the design work is making sure the setup respects them.
VoIPPBXSIPSpeech-to-textSentiment analysisCall analytics

MSP and IT automation

For managed service providers and enterprise IT teams, I automate the work between tools: connecting RMM and monitoring platforms with MSP systems, CRM and API workflows, Microsoft 365, Active Directory and networking. The goal is fewer repetitive tickets and consistent, repeatable operations.

What I build

  • Tool-to-tool integrations between RMM, monitoring, PSA/ticketing and CRM.
  • Alert handling that enriches, de-duplicates and routes alerts instead of flooding a queue.
  • User lifecycle automation for onboarding and offboarding across Microsoft 365 and Active Directory.
  • Standardised, repeatable procedures so common tasks are done the same way every time.

What changes for your team

  • Fewer repetitive tickets reaching engineers.
  • Less copy-pasting between consoles.
  • Consistent results across clients and sites.
  • A written trail of what each automation does.
RMMMonitoringMicrosoft 365Active DirectoryNetworkingCRMAPI workflows

Algorithmic trading automation

My trading work covers market-data pipelines and options-chain analysis that connect broker APIs with strategy backtesting and real-time trading signals. It is engineering work on systems and data, not a promise of returns.

What I build

  • Market-data pipelines that collect and store live and historical data.
  • Options-chain analysis tooling for exploring strikes, open interest and changes over time.
  • Backtesting frameworks to test a strategy's rules against historical data.
  • Real-time signals and alerts delivered to dashboards or messaging apps.
  • Broker API integration for data and, where appropriate, order workflows.

Principles

  • Reliable data and clear logs come before clever strategy logic.
  • Backtests are only as good as their assumptions; I document them.
  • Risk controls and manual overrides are built in from the start.
Trading involves risk of loss. Nothing here is investment advice or a guarantee of returns; I build the tools and systems, and the trading decisions remain yours.

You can see some of the tools I use day to day in the workspace.

Broker APIsOptions chainBacktestingReal-time signalsPythonData pipelines

How an engagement works

  1. Discovery callWe talk through the workflow, the systems involved and what a good outcome looks like.
  2. Scope and planA clear written scope: what will be built, what it connects to, and how we will know it works.
  3. Build and reviewI build in small steps and show working results early, so feedback arrives before the end.
  4. Deploy and hand overThe solution goes live with logging and documentation, so your team can run and understand it.
  5. Support and iterateFixes and improvements as the real-world workflow shows what needs adjusting.

Common questions

Do I need to replace my existing systems?

No. The approach is to integrate with what you already run: your phone system, ticketing, CRM, databases and APIs.

Can you work with our own data without it going to a public service?

That is a design question we settle early. Depending on your requirements, the options include using providers with suitable data terms, limiting what is sent, or hosting components yourself.

What do you need from us to start?

A description of the process you want to improve, access to the relevant systems or sample data, and someone who knows how the work is done today.

Is the trading work financial advice?

No. It is software and data engineering. I do not manage money or recommend trades.

How do I get started?

Book a consultation and describe the workflow you have in mind. We will take it from there.

Have a workflow in mind?

Tell me about the process that eats your team's time and we will work out whether and how to automate it.

Schedule a consultation Back to home