Callan Jackson
Junior Backend / Software Engineer · Python · Django · C#/.NET · Azure
Profile
Backend-focused software engineer with a First-Class Honours BSc in Computer Science (University of London / Goldsmiths), earned while working full time as cabin crew.
I build production-shaped applications and deploy them for real: four of my projects run live on Azure, including an FCA-compliant pension analysis platform in C#/.NET 10 and an AI-powered incident root-cause and remediation agent (Azure OpenAI, Semantic Kernel) and a containerised music-recommendation platform with real-time WebSocket search. Day to day that means Python and Django backed by PostgreSQL, Redis, Celery and Docker, plus C#/.NET on Azure with infrastructure as code and CI/CD.
I'm looking for a junior backend or software engineering role where I can keep shipping reliable, well-tested services. Comfortable under pressure, detail-driven, and used to high-compliance, safety-critical environments.
Technical Skills
- Languages: Python, C#, TypeScript / JavaScript, SQL.
- Backend & frameworks: Django, Django REST Framework, .NET 8 & 10 / ASP.NET Core, Blazor, Django Channels (WebSockets), Celery, Node.js / Express, React 19.
- Databases & caching: PostgreSQL, MySQL, Redis, DuckDB.
- Cloud & infrastructure: Azure (App Service, Azure SQL, Key Vault, Application Insights, Azure OpenAI, Bicep IaC), Docker & docker-compose, GitHub Actions CI/CD with OIDC federated credentials, Git & GitHub, REST API design, Linux, OpenAPI.
- Data & ML: pandas, NumPy, scikit-learn, TensorFlow / Keras, Matplotlib, Seaborn.
- Concepts: OOP, database design & normalisation, authentication & role-based permissions, async / event-driven design, LLM integration & AI-agent workflows (Semantic Kernel), AI-assisted development with review discipline, decimal precision & numerical methods, content-based recommendation.
Projects
SwitchPoint: FCA-Compliant Pension Switching & DB Transfer Analysis Platform (live on Azure)
C# · .NET 10 · React 19 · TypeScript · EF Core · Azure SQL · QuestPDF · Bicep · Docker
- Deterministic calculation engines for UK pension advice, every rule traced to the FCA Handbook section it implements: critical yield and reduction in yield solved numerically with Brent’s method, the COBS 13 effect-of-charges table, a COBS 19 Transfer Value Comparator over a documented mortality basis, UK income tax and allowances, and Monte Carlo cashflow modelling with Cholesky-correlated returns.
- All money and rate arithmetic in
decimal, including hand-written Exp, Ln and NthRoot, so a suitability report reproduces to the penny on any platform — verified by closed-form oracles, inverse round-trips and FsCheck property tests across 366 tests.
- Tamper-evident per-firm SHA-256 audit chain with an append-only interceptor; suitability reports rendered to PDF and Word, each pinned to the analysis version and result hash it was produced from.
- React SPA and API served from one container on Azure App Service, provisioned with Bicep and deployed by GitHub Actions using OIDC federated credentials.
IncidentPilot: Automated Cloud Incident Root-Cause & Remediation Agent (live on Azure)
C# · .NET 8 · Azure OpenAI · Semantic Kernel · Blazor · EF Core · Bicep · GitHub Actions
- End-to-end diagnostic pipeline: failures are ingested via API, fingerprinted to correlate recurrences, then analysed by a two-stage AI workflow (Semantic Kernel over Azure OpenAI) that returns a schema-validated root-cause verdict with confidence, evidence and blast radius, plus a remediation plan and auto-generated pull request.
- Deterministic rule-based fallback engine so diagnosis survives LLM outages or malformed output; live Blazor dashboard; 31 unit and integration tests.
- Provisioned with Bicep IaC and deployed continuously via GitHub Actions using OIDC federated credentials, with no stored secrets.
NextTrack: Content-Based Music Recommendation Platform (Final Year Project · live on Azure)
Python · Django · Django Channels · PostgreSQL · Redis · Celery · librosa · Docker · NumPy
- Full-stack, containerised web app recommending tracks by acoustic similarity using a 5-dimensional audio-feature vector and Euclidean distance, with an adaptive centroid that learns from in-session like/dislike feedback.
- When Spotify deprecated the audio-features API the system depended on, replaced it with an in-house DSP pipeline: librosa feature extraction (beat tracking, harmonic/percussive separation, key detection) over 30-second preview clips, run asynchronously on Celery with results streamed back into the UI as each track completes.
- Hybrid search combining a local catalogue with live Deezer queries over real-time WebSockets, plus in-browser playback of the exact audio the recommender measured.
- Integrated external metadata services (MusicBrainz, Wikidata, Last.fm) with graceful fallbacks; deployed to Azure with a production Docker Compose stack (PostgreSQL with pgbouncer, Redis, prioritised Celery queues) behind automatic HTTPS, with GitHub Actions CI.
Lernova: Multi-Role Learning Platform (live on Azure)
Python · Django 4.2 · Django REST Framework · Django Channels · Celery · PostgreSQL · Docker
- Role-based platform (student / teacher / admin) with course management, enrolment with capacity limits via atomic transactions, validated file uploads, and real-time per-course chat over authenticated WebSockets.
- Security-hardened following a full audit: per-room WebSocket authorisation, CSRF-safe state changes, HSTS and hardened headers, API throttling and login lockout; upgraded Django 3.2 → 4.2 LTS.
- Custom DRF permission classes and REST API; deployed to Azure with Docker Compose behind automatic HTTPS.
easyJet Ops Intelligence: UK Aviation Punctuality Analytics
Python · DuckDB · scikit-learn · pandas
- Analytics pipeline exploring UK flight-punctuality data with DuckDB and machine learning, combining knowledge of the airline industry with data engineering.
Evolving Mountain-Climbing Creatures: Genetic Algorithms
Python · PyBullet · NumPy
- Evolutionary simulation evolving virtual creatures to climb procedurally generated terrain in a physics engine.
- Designed a custom stability-based fitness function to prevent reward-hacking; robust parallel evaluation (ProcessPoolExecutor with per-creature timeouts) for multi-hour runs.
Experience
Cabin Crew, easyJetLondon Gatwick · Feb 2023 – Present
- Deliver safety-critical procedures and customer service to hundreds of passengers per shift in a regulated, high-compliance environment.
- Work effectively under time pressure within rotating crews; recognised for reliability and attention to detail.
- Earned a First-Class Honours BSc in Computer Science alongside full-time shift work, four years of sustained self-directed study and time management.
Sales & Installation, Soul Shutters2020 – Feb 2023
- Customer consultation and sales of bespoke window shutters; precise on-site measuring, fitting and installation requiring high accuracy.
Earlier roles2015 – 2020
- A range of customer-facing and operational roles across retail, hospitality and real estate.
Education
BSc Computer Science, University of LondonApr 2022 – Apr 2026
Academic direction by Goldsmiths, University of London · First-Class Honours
Relevant modules: Advanced Web Development, Databases & Advanced Data Techniques, Machine Learning & Neural Networks, Data Science, Artificial Intelligence, Algorithms & Data Structures, Software Design & Programming.