Jonathan Franklin
About
I’m Jonathan Franklin, a senior software engineer and technology leader with 20+ years of commercial development experience. I specialise in Python, AWS, data engineering, AI/LLM systems and algorithmic trading.
At Pricesearcher, I am Chief Technology Officer and Algorithmic Trader, responsible for the technology behind a price-comparison platform serving multiple European markets and for developing and operating the company’s automated FX/CFD trading strategies.
I work hands-on across the full lifecycle: research, architecture, implementation, deployment, monitoring and ongoing operation. My background also includes CTO, head-of-technology and technical-leadership roles in startups, contract teams and established engineering organisations.
Experience
Chief Technology Officer / Algorithmic Trader · Pricesearcher
My work on the price-comparison business has covered merchant feed ingestion, Spark-based product enrichment, Elasticsearch indexing, multi-source search APIs and the React/Next.js website. I have also worked on affiliate integrations, advertising and revenue reporting, and the Airflow and AWS infrastructure supporting the data platform.
My trading work spans MQL5 Expert Advisors, Python execution engines, an event-driven MetaTrader bridge, market-data pipelines, strategy research, backtesting, risk analysis and live reporting. I have also built local LLM-powered strategy research workflows and an MCP-connected analytics assistant.
Chief Technology Officer & Co-Founder · Nimbli
I co-founded Nimbli and led its technology strategy, building Python backend APIs and mobile applications for financial services aimed at the unbanked population in East Africa, alongside Kubernetes infrastructure on Google Cloud Platform.
Head of Technology (Contract) · Fusepump
I took responsibility for a team of approximately ten developers and testers, covering delivery, recruitment, team development, technical strategy and the company's AWS infrastructure.
Software engineering roles
I worked across eCommerce, gaming, telecoms and automotive projects at Salmon, Gamesys, Virgin Media, Honda, Gamma Telecom and Intersoft, using Java and PHP and leading delivery on parts of major customer-facing systems.
Technical focus
- Backend systems and APIs. Python services using Flask and FastAPI, third-party API integration, asynchronous processing, Redis caching and SQL-backed applications. The product search API is an example of bringing several providers together for different client applications.
- Data engineering and search. Feed ingestion, data normalisation, Spark enrichment, Parquet datasets and Elasticsearch indexes, orchestrated with Airflow and AWS services. My catalogue pipeline work connects raw merchant feeds to the product data used by search.
- Web applications. React, TypeScript and Next.js for public-facing search experiences and internal reporting tools. This includes server-side rendering, configurable page layouts, affiliate tracking and deployment.
- Infrastructure and operations. AWS, Docker, Terraform, Pulumi and CI/CD with Jenkins and GitHub Actions. My work includes migrations, scheduled workloads, data-freshness checks, alerting and adjustments to infrastructure capacity and job frequency.
- Trading systems and applied machine learning. Python and MQL5 for event-driven execution, strategy configuration and risk controls; PyTorch, scikit-learn and LightGBM for research and model-based strategies. This includes statistical arbitrage, historical replay, fee-aware evaluation and position-sizing models. I also develop Quantrade, a personal research system for comparing models and trading ideas.
- AI/LLM engineering and AI-assisted development. Self-hosted models with llama.cpp and ninfer, cloud models where required, specification-led agentic workflows and MCP integrations that connect agents to project and trading data.
How I approach the work
I pay attention to the boundaries between systems: inconsistent provider responses, timestamps that do not line up, country-specific data ending up in the wrong table, and configuration changes that affect running strategies. Those details determine whether the result is useful, not just whether the code runs.
I prefer to build on an existing platform where it fits. That can mean adapting an open-source image proxy, extending a shared product pipeline to another market, or using partitioned files on S3 for date-range retrieval rather than introducing a database.
I use local and cloud LLMs to accelerate research, challenge requirements, plan changes and implement them, while retaining engineering oversight of architecture, testing and production quality.
In trading research, I distinguish a promising backtest from evidence that a strategy will hold up in live use. Data timing, transaction costs, drawdown and execution assumptions matter as much as the model.
Contact
I’m open to consulting engagements and full-time roles involving senior hands-on engineering or technical leadership, particularly across backend systems, data platforms, AI/LLM tooling and algorithmic trading. Use the contact form to tell me about the problem, the existing system and the help you need. You can also find me on GitHub and LinkedIn.