Services
What I build.
Data science, analysis and model training first, then everything it takes to put the result to work — dashboards, deep learning, AI solutions, AI automations, and the full-stack product around them. Ten service lines that combine into complete builds.
Data science first
Trained models, not borrowed benchmarks.
My degree is in Data Science and it is still the centre of the work: understanding the data, reporting what it says, engineering the features, training the model, and proving it holds before anything ships.
- 01Data & analysisProfiling, cleaning and exploratory analysis — finding the leakage, the imbalance and the signal. Often this stage is the deliverable: a report and a dashboard, no model required.
- 02Feature engineeringTurning raw operational data into features a model can learn from, with the transformations reproducible in a pipeline.
- 03Model trainingClassical ML and deep learning, trained on my own GPU — architecture choices, hyperparameter tuning and tracked experiments.
- 04EvaluationHeld-out validation, the metrics that match the business cost of being wrong, and error analysis on the cases that failed.
- 05DeploymentThe model behind an inference API, containerised, monitored, and wired into the product or the automation that uses it.
Data analysis
Digging through the data to answer the question actually being asked — trends, drivers, segments and the caveats that come with them.
Dashboards & BI
Power BI, Tableau, Looker Studio, or Python with Matplotlib, Seaborn, Plotly and Streamlit — whichever your team will actually keep using.
Model training
Classification, regression, forecasting and vision models trained from scratch or fine-tuned from a strong checkpoint.
Deep learning
CNNs, RNN/LSTM and transformer architectures in PyTorch and TensorFlow, including transfer learning and quantised inference.
AI solutions
Complete AI products — computer vision pipelines, RAG assistants and agent systems built around the model and its data.
AI automation
n8n and agent workflows that put the model to work: documents read, records enriched, reports written, CRMs kept current.
Data Science & Machine Learning
The core of what I do: get inside the data, find what actually predicts the outcome, and train a model that holds up on inputs it has never seen.
- Exploratory Data Analysis
- Feature Engineering
- Classification
- Regression
- Forecasting
- Clustering
- Statistical Analysis
- Model Evaluation
- Experiment Tracking
- Typically from
- $450USD
- Timeline
- 1 — 3 weeks
Data Analysis & Dashboards
Making the data say something useful — analysis that answers the question actually being asked, and dashboards a team will still be opening in six months.
- Exploratory Data Analysis
- Power BI
- Tableau
- Looker Studio
- Matplotlib
- Seaborn
- Plotly
- Streamlit
- Excel / Google Sheets
- KPI & Metric Design
- Automated Reporting
- Typically from
- $300USD
- Timeline
- 3 days — 2 weeks
Deep Learning & Model Training
Neural networks trained properly — architecture selection, transfer learning, fine-tuning and long training runs on my own GPU, then packaged behind an inference API you can call.
- PyTorch
- TensorFlow
- CNNs
- RNN / LSTM
- Transformers
- Transfer Learning
- Fine-Tuning
- Hyperparameter Tuning
- GPU Training
- Inference APIs
- Typically from
- $550USD
- Timeline
- 1 — 4 weeks
Computer Vision
Vision systems that read the real world: detection and tracking, document and face analysis, and the fraud signals that sit on top of them.
- Object Detection
- YOLO
- OCR
- Face Analysis
- ID Verification
- Spoof Detection
- Deepfake Detection
- Video Analysis
- Typically from
- $600USD
- Timeline
- 2 — 5 weeks
LLM & Generative AI
Applications built on language models — retrieval-augmented assistants, agents and tool use, grounded in your data rather than model recall.
- LLM Applications
- RAG
- AI Assistants
- AI Agents
- LangChain
- LangGraph
- MCP Servers
- Local LLMs
- Typically from
- $500USD
- Timeline
- 1 — 4 weeks
AI Automation & Agents
Workflows that remove the manual middle step — email, CRM, content and publishing operations wired together and run by agents.
- n8n
- AI Agents
- API Automation
- CRM Automation
- Email Automation
- Content Automation
- Multi-step Workflows
- Typically from
- $300USD
- Timeline
- 3 days — 2 weeks
Data Engineering & Pipelines
The layer underneath the analysis: pipelines and warehouses that turn scattered operational data into something a dashboard — or a model — can trust.
- ETL / ELT
- Data Warehouses
- SQL
- SSIS
- Data Pipelines
- Data Modeling
- Data Quality
- Scheduled Jobs
- Typically from
- $400USD
- Timeline
- 1 — 4 weeks
Backend Development
APIs and services designed to be operated: clear contracts, real validation, sensible auth and jobs that survive failure.
- Python
- FastAPI
- Django
- REST APIs
- PostgreSQL
- MySQL
- MongoDB
- Authentication
- Background Jobs
- Typically from
- $350USD
- Timeline
- 1 — 4 weeks
Full-Stack Development
Complete products — the model, the API and the interface people actually use — shipped as one coherent application.
- React
- Next.js
- TypeScript
- Python Backends
- Dashboards
- SaaS Applications
- AI-Powered Web Apps
- Typically from
- $600USD
- Timeline
- 2 — 6 weeks
DevOps & Integration
Getting it live and keeping it there: containers, pipelines, deployments and the integrations that connect your stack together.
- Docker
- GitHub
- CI/CD
- API Integrations
- Deployment
- Infrastructure Automation
- Typically from
- $250USD
- Timeline
- 3 days — 2 weeks
Prices are indicative starting points for a scoped piece of work. A firm quote follows a short scoping call — no obligation.
Engagement
Three ways to work together.
Pick the shape that fits the problem — none of them lock you into the wrong one.
Hourly
$12/ hour
For exploratory work, ongoing support and anything where the scope is still moving. Tracked and reported honestly.
- No minimum commitment
- Time tracked and reported
- Ideal for iterations and support
Fixed project
Scoped
For a defined deliverable with a clear finish line. Scoped up front, priced once, delivered against a written spec.
- Indicative from-prices per service
- Scoped and quoted before work starts
- Milestone-based delivery
Monthly retainer
Scoped
For teams who need consistent AI and automation capacity month to month rather than one-off builds.
- Reserved weekly capacity
- Priority turnaround
- Rolling backlog, monthly review
How I work
Four steps from problem to production.
No black boxes. You see the reasoning, the architecture and the trade-offs at every stage.
Understand
Understand the business problem, the requirements, the users and the constraints — before a single line of code is written.
Design
Design the architecture, the workflow, the data pipeline and the AI approach, with the trade-offs made explicit.
Build
Develop, integrate, test and iterate — in visible increments rather than one long silence.
Deploy
Ship the solution, document how it works, and stay available for the questions that come after launch.
Infrastructure
Model training runs on my own hardware
An RTX 3080 Ti with 12 GB of VRAM sits on my desk, which means training, fine-tuning and long experimental runs happen locally — without renting cloud GPUs by the hour or passing that cost on to you.
- Local training and fine-tuning without hourly cloud GPU cost
- Fast iteration on detection, vision and deep learning models
- Local LLM inference and quantised model evaluation
- Cloud training still available when a job genuinely needs more
You own everything
Code, models, weights, workflows and documentation are handed over. No lock-in, no hosting hostage.
Visible increments
You see working pieces as they land rather than a long silence followed by a big reveal.
Written down
Architecture decisions, trade-offs and how to run it are documented — so the next engineer is not stuck.
Work with me
Hire me where it suits you.
Same engineer, same standard — pick the platform that fits how your team contracts.
Upwork
Hire me for AI engineering, automation, LLM, data and full-stack projects — contract-protected, milestone based.
View Upwork ProfileFiverr
Explore packaged AI, automation, LLM and development services with a defined scope and turnaround.
View Fiverr ProfileDirect
Prefer to skip the platforms? Message me on WhatsApp and we can scope it in a single conversation.
Chat on WhatsAppLet’s build
Have an idea that needs AI?
Whether you need an AI model, an automation workflow, a data pipeline, a backend system or a complete AI-powered product — let's build it.