About Ali
AI Engineer with 7+ years of experience delivering end-to-end ML/AI solutions from problem framing through production deployment and monitoring. Proven track record building personalisation engines, customer analytics models (churn, propensity, segmentation), and optimisation systems across energy, retail, and services sectors. Deep expertise in Python, LLMs/GenAI (RAG, prompt engineering, multi-agent systems), classical ML (time-series forecasting, classification, NLP), and modern MLOps practices on Azure and AWS. Cambridge MPhil, First-Class Honours.
English
Native or bilingual
Can work on-site
London (up to 50km)
Experience
- JMAN GroupSenior AI Engineer | Customer Analytics & Personalisation NLP-LLM PipelineDecember 2025 - February 2026 (2 months)
- Built End-to-End Pipeline: Developed a customer analytics pipeline for multilingual European markets (call transcripts, chat, and voice logs) to enable data-driven segmentation and personalised service strategies.
- Advanced NLP Workflows: Developed BERTopic workflows using HDBSCAN clustering and UMAP to identify distinct customer segments and sentiment patterns.
- LLM Integration: Orchestrated GPT-4o-mini for multilingual translation and hierarchical topic labeling.
- Cost Efficiency: Reduced API costs by 50% through intelligent batching and optimised processing.
- Semantic Analysis: Engineered memory-optimised multilingual embeddings for processing large (GB+) datasets.
- Production Engineering: Designed a YAML-based config-driven architecture for rapid experimentation and implemented comprehensive logging and monitoring.
- PropTech & Geospatial IntelligenceStartup PropTech - AI EngineerOctober 2025 - December 2025 (2 months)
- Architected and developed a production-ready multi-agent AI orchestration platform using LangGraph and LangChain to automate property development due diligence workflows, reducing manual analysis time by 85% and enabling scalable batch processing of land title assessments.
- AI Agent Development & Orchestration: Designed and implemented a modular multi-agent system with 7 specialised AI agents (Legal Due Diligence, Planning & Regulatory Compliance, Flood Risk Assessment, Protected Land Analysis) using LangGraph state machine architecture with gate-based workflows, conditional branching, and early stopping logic to optimize computational efficiency.
- Workflow Automation: Built config-driven agent architecture with YAML-based configuration management, enabling dynamic agent behavior modification without code changes. Implemented advanced workflow automation with state persistence, error recovery mechanisms, and comprehensive audit trails for production reliability
- Technical Implementation: Engineered memory-efficient chunked processing for large-scale spatial datasets (GB+), implementing row-group-based Parquet processing with progress tracking, reducing memory footprint by 90% while maintaining sub-minute processing times. Integrated multiple government data sources (Historic England, Environment Agency, Natural England) with 400K+ spatial features, converting datasets to optimised Parquet format achieving 67% file size reduction and 5-10x faster query performance.
- Production Engineering: Created batch processing capabilities with CLI interface supporting single-title, batch-file, and interactive modes, enabling processing of hundreds of titles with comprehensive error tracking. Established modular architecture with shared common utilities, reducing code duplication by 60% and enabling rapid agent development.
- HarnhamGlobal Data & AI recruitment specialistsAugust 2025 - October 2025 (2 months)
- Architected and implemented a modular data pipeline to handle end-to-end processing of complex sales data for global FMCG clients, which is a critical first step for any large-scale AI/ML initiative.
- Utilized the Polars library and its LazyFrame API to build a memory-efficient and performant data processing engine, optimising the handling of large datasets often required for training and fine-tuning AI models.
- Championed advanced data engineering principles to create a robust, production-ready solution. The modular design, comprehensive error handling, and use of configurable settings ensure the pipeline is reliable and easy to maintain.
- Implemented extensive unit and integration testing with pytest, leveraging fixtures and parametrisation to validate complex data transformations, edge cases, and error scenarios, maintaining high code quality and reliability standards.
Recommendations
Be the first to recommend Ali
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- MPhilUniversity of Cambridge, Emmanuel College2017MPhil
- BEngUniversity of Southampton2016BEng