Benchmarking and perf team. Internal tooling.

Boutique algorithmic and prop trading firm targeting Eastern Europe, India, South Africa. I built two versions of their main algorithmic trading engine: experimental in Common Lisp and one in production in Erlang/OTP. All trading agents modelled as explicit finite state machines driven by events. Wrote a few external components in Python, R and Julia. Preped their infra for trading on NSE, India including exchange colo resources.

Conversational AI agents inside Telegram

• WIP, working towards MVP

AI-powered Telegram bot that turns your study materials into bite-sized drills aka Anki flashcards and helps you memorize anything and learn faster. Way faster. ~1KLOC of TypeScript and ~6KLOC of Clojure. Integrations with Telegram and LLMs are all custom written.

Made tech blogging as simple as creating a Github gist. My hoots can be found at vlad.git.ht

Mostly one-off engagements with unhealthy mix of Python and JavaScript with Nodejs.

Revenue-based financing and SME advance payments for the XXI century. JVM on the backend and ClojureScript with React on the client. I did half of both as well as comlpete OpenBanking and Accounting integrations for the UK and US. Integration with data-science and product teams, internal ledger. Deployed to AWS with Terraform and Kubernetes, managed MySQL. Occasional Python for internal tooling and AWS lambdas.

Infrastructure for Chia (XCH) cryptocurrency and blockchain as a service. On-demand "plotting" and "farming" rigs. Mostly Python for infrastructure management and core functionality, cloud-init and Ansible for provisioning, customized SmartOS hypervisor for virtualisation and datacenter management. Node.js with Express.js on the server and a splash of React on the client. Occasional smart-contract hacking with Chia Lisp.

Started out as a Clojure SWAT team targeting JVM then branched out into .NET with Clojure on CLR, C# and F#. We mostly contract with FinTech startups and financial institutions. We prototype, MVP, take projects to scale out, hire, on-board and hand off to staff engineers, port and maintain existing JVM and .NET products. We recently built a Funding Circle competitor in under 5 months and lauched it in UK and US. I personally implemented the entire Open Banking journey, integration with payment processing providers, etc.

Three of us built a cognitive assistant for investment banking researchers. Mostly Clojure, ClojureScript. C# and F# with LINQ, Elastic Search, Kafka, MongoDB to integrate customer-side data sources. We also did much of the machine learning and AI hacking: Recursive Neural Networks (pre-LLM approaches), linguistic models and NLP, Lucene, etc. Kubernetes for infrastructure. Java Swing GUIs for internal ML tools.

Put together Droit's first London engineering team. All Clojure and ClojureScript, handpicked and bright. Still proud. Otherwise same hot fintech startup hacking as before the move to London.

Built an expert system for compliant trading. Sneaked Clojure into a few unsuspecting giants like Goldman, Morgan Stanley, etc. You'd typically find me designing DSLs, implementing compilers, parsers, rules-based engines, putting together simple browser-based GUIs and whatever else the startup life would have me do.

Officially a member of the Search Interfaces Development Infrastructure group, but mostly I wrote backend tools for source to source compilation - engines to write your template engines. If I was lucky and did it right frontend developers would get to use my work and take all the credit. Mostly JavaScript and Node.js.

Sales and trading mostly equity options with occasional structuring here and there.

EM derivatives structuring and sales. Mostly rates and FX with an occasional commodities trade.