Tanveer Hoque
Applied AI engineer building production LLM systems, with the backend discipline that keeps them cheap, fast and safe
Cardiff, United Kingdom | tanvoid0@gmail.com | linkedin.com/in/tanvoid0
Professional Summary
Applied AI engineer in Cardiff, working on the part of AI that has to survive production: what it costs, how long it takes, and what a model is allowed to reach. I wrote agent-platform, a Rust service putting seven hosted model providers and two local runtimes behind one OpenAI-compatible endpoint, forwarding each caller’s own key rather than storing any. I wrote claude-code-plugins, an agent spend limit and routing layer measured 68% cheaper over four identical tasks and replayed across 395 real sessions before shipping. Underneath is four years of production backend engineering, Java and Spring Boot finance microservices on PostgreSQL, RabbitMQ, Google Cloud and Jenkins, and that is where the habits come from: fix the query before reaching for a cache, and never publish a number nobody measured.
Key Skills
AI, models & agents: Claude Code, Cursor, Ollama, OpenAI-compatible APIs
Languages: Java, TypeScript, SQL, JavaScript, Rust, Python, Bash, C++
Frameworks: Spring Boot, Spring Batch, JPA / Hibernate, Node.js, NestJS, FastAPI, Axum / Tokio, GraphQL, React, Next.js, Astro
Data & messaging: PostgreSQL, RabbitMQ, Liquibase, MongoDB
Delivery & operations: Jenkins, Docker, Kubernetes, Google Cloud, Vercel, GitHub Actions
Testing & tooling: JUnit, Git, Jira
Work Experience
Software Engineer | Apr 2023 – Present
Civica Ltd. (formerly OLM Systems Ltd.), United Kingdom
- Critical finance API latency: 99s → 0.9s.
- Payment processing: event-driven over RabbitMQ.
- Release path: manual steps → Jenkins pipeline.
- Build and maintain Spring Boot microservices for finance, with JPA queries over PostgreSQL and schema managed by Liquibase.
- Run the payment flow as an event-based process over RabbitMQ, including the scheduling and the failure paths.
- Deploy and operate the services on Google Cloud, day to day, alongside the Jenkins pipelines that build them.
- Stack: Java, Spring Boot, JPA, Hibernate, PostgreSQL, Liquibase, RabbitMQ, GraphQL, Node.js, React, Google Cloud, Jenkins, Docker.
Junior Software Developer | Jan 2022 – Apr 2023
OLM Systems Ltd., United Kingdom
- Test coverage: automated suites in pre-release.
- Delivered features across Spring Boot services and their React front ends.
- Improved test coverage for microservices with JUnit in the pre-release environment.
- Took part in story estimation and planning, and kept the reference notes the team used for both.
- Stack: Java, Spring Boot, JPA, PostgreSQL, JUnit, React.
Intern | Jan 2019 – Jun 2019
Samsung Research & Development Institute Bangladesh, Dhaka, Bangladesh
- Worked through algorithm and data-structure problems against the standards of a large product organisation.
- Took part in code review and problem-solving sessions with engineers on shipping teams.
- Stack: C++, Java, Algorithms.
Projects
agent-platform | Author | 2026 to present
Rust, Axum, Tokio, sqlx, SQLite, iced, Docker, Ollama, LM Studio, Python — https://github.com/tanvoid0/agent-platform
- Nine upstreams behind a single endpoint, seven hosted providers and two local runtimes, with no provider-specific path in the client.
- A prompt can be answered by a model on the same machine: no key issued, nothing leaving the host, the same code as the hosted route.
- A caller’s own provider key is accepted per request, and the base URL it may be pointed at is allowlisted, https-only and never a raw IP, so BYOK cannot be turned into a way to reach internal services.
crack-detector | Author | 2026
Java, HTML, PowerShell
- Sessions survive a refresh: Postgres 17 with Flyway migrations and optimistic locking behind JPA.
- The state machine is covered by an integration test and the controller by a @WebMvcTest slice, with Testcontainers standing up the database.
- Actuator health checks cover both Postgres and the LLM, so a laptop demo comes up in under two minutes on Docker Compose.
claude-code-plugins | Author | 2026 to present
Node.js, Claude Code, Hooks, Subagents, Opus / Sonnet / Haiku — https://github.com/tanvoid0/claude-code-plugins
- Four identical tasks, routed against unrouted: 298,615 tokens versus 921,297, and $0.40 versus $1.28, so 68% cheaper with every result verified by execution rather than taken from the agents’ own reports.
- The gate replayed over 395 real sessions (27.7M output tokens, $690 of spend) would have stopped 6.1% of them, and no others.
- A dispatcher given only the six published descriptions and no repo access routed 18 of 18 tasks correctly, and still 18 of 18 after the descriptions were compressed 31%.
Education
MSc Computer Science | Sep 2020 – Sep 2021
University of South Wales, Cardiff, United Kingdom
Dissertation: implementing a secure lost & found application.
BSc Computer Science | Jan 2015 – Jan 2019
Daffodil International University, Dhaka, Bangladesh
Dissertation: data analysis on home forestation using IoT. Robotics team and competitive programming.
Languages
English: Full professional; Bengali: Native; Hindi: Conversational; Urdu: Conversational
Additional Information
Work authorisation: Authorised to work in the United Kingdom
Notice period: One month
Availability: Open to Applied AI, AI platform and backend engineering roles in the UK. Remote or hybrid.
GitHub: github.com/tanvoid0