alex → cv

• Built a DAG-based orchestration framework that lets autonomous agents carry out multi-step data analysis end to end
• Designed a DSL that non-technical users write analysis workflows in, mixing LLM agents with conventional ML models in the same pipeline
• Extended the semantic data model with ontological abstractions and higher-order business concepts — the layer that makes automated root-cause analysis and data discovery possible at all
• Built the feedback loop that turns user corrections into evaluation data and downstream few-shot examples, so the system improves without a retraining cycle
• Wrote tabular insight agents on Jinja templates that carry full lineage and provenance, so any generated number can be traced back to source for enterprise audit
• Ran a fault-tolerant distributed worker fleet on AWS (ECS/SNS/SQS) with dead-letter queue handling and zero-downtime deploys
• Lead AI engineering for the analytics product, turning analyst workflows that were manual or simply not possible into automated ones

• Started the AI function from nothing and shipped its first systems to production
• Built the RAG pipeline behind the product's generated responses, covering retrieval, chunking, and grounding
• Designed the Postgres schema and backend on Supabase, including authentication and access control
• Architected the API layer the AI pipelines sit behind, which cut latency and raised throughput
• Moved infrastructure to CloudFormation and CI/CD to GitHub Actions
• Added monitoring across the AI pipelines to catch quality regressions before customers did
• Built ontological models that gave the data model a consistent vocabulary to work from
• Restructured the Postgres schema as the access patterns became clear and load grew
• Coordinated with external development teams on platform integration, and ran project management out of GitHub

• Advise startups and established companies on where AI belongs in their stack, and where it does not
• Delivered a keynote on ChatGPT to a room of CIOs and founders
• Advised a blockchain unicorn on engineering process and operational flow
• Run technology assessments and build-versus-buy analysis for teams committing to an AI direction
• Led migrations to AI-integrated systems, reducing cost and manual effort
• Consulted on personal knowledge management systems built around LLMs
• Help early-stage startups choose a technology stack they will not have to abandon in a year

• Led a cross-functional engineering team, working alongside the music, research, and sales sides of the company
• Ran GCP infrastructure: Kubernetes and Istio, monitored with Grafana and Prometheus
• Rebuilt CI/CD on Jenkins with Cypress end-to-end coverage
• Built a custom data ingestion pipeline and automated the manual steps that had grown around it
• Introduced Scrum, along with the scheduling and coordination practice needed to make it stick
• Set technical direction against customer and partner requirements, planning releases with the founders
• Mentored engineers on the team and ran training sessions off the back of operational reviews

• Took the product from concept to launch: a cloud platform that emulates supercomputer environments so teams can develop and test at scale without waiting for time on the real machine
• Raised early rounds from government, VC, and angel investors
• Won early customers including Fortune 500 companies and national governments
• Grew the team from 3 to 15 in the first year
• Set the business model and go-to-market strategy, with the financial model underneath it
• Managed relationships with several levels of government across the region
• Ran investor relations and board communications
• Negotiated the contracts with partners and suppliers

• Co-founded Archanan out of the programme
• Secured letters of intent from early customers before committing to a build
• Tested several ideas against the market and killed the ones that did not hold up
• Built the financial model and the go-to-market plan that came out of it
• Led the first fundraise
• Formed the early partnerships the company ran on

• Designed and built the MVPs, then the production cloud middleware that replaced them
• Directed algorithm development for the core computational problems the product depended on
• Raised seed capital from VCs and angels
• Led the engineering team and set the code review and QA standards it worked to
• Got the company into the Grand Central Tech accelerator
• Worked directly with clients on what to build next

• Designed, prototyped, and tested audio synthesizers, including the circuit design
• Led seminars on where music and mathematics meet
• Turned research into work that could actually be performed and heard
• Collaborated with people from the music technology industry on novel audio hardware

• Researched optimal wavelet bases for audio compression, searching for a general procedure rather than a one-off basis
• Ran a supercomputing project funded by the High Performance Computing Consortium of New York
• Performed real-time signal analysis on spectrum data
• Kept the project documentation and datasets in a state other researchers could pick up

First software job, and first hardware job — where I learned how shipped products actually get made.

• Wrote production firmware for digital guitar pedals
• Shipped feature releases through the team's existing continuous integration process
• Worked to the established test practice, catching firmware defects before release
• Wrote the technical documentation and user manuals
• Fed customer feedback back into what got built next

• Assembled and tested printed circuit boards for audio processing units
• Ran technical customer service, including repairs and returns
• Serviced customer hardware sent back to the shop
• Trained new staff on assembly and quality control