Backend Software Engineer —
Building Java/Spring Boot microservices and payment APIs for banking platforms at HSBC across 4 global markets — plus GitHub Copilot AI agents that automate Java upgrades and security fixes.
I'm a Backend Software Engineer at HSBC Technology India, building distributed systems that power banking, card processing and payment workflows across India, Singapore, UAE and the UK.
My craft is Java/Spring Boot microservices — API orchestration flows (PAPI), service-to-service security with OAuth 2.0/JWT, reactive programming with RxJava, and shipping on Kubernetes in cloud-native environments.
At HSBC I also build GitHub Copilot AI agents that migrate Java APIs from JDK 8 to JDK 21 and fix critical open-source dependency vulnerabilities. Beyond work, I explore GenAI engineering — agentic workflows, RAG pipelines and autonomous code-generation systems using LangChain and frontier LLMs. I also hold a Diploma in Data Science & Programming from IIT Madras, completed part-time alongside my HSBC role.
Built the backend for Visa Click-to-Pay in-app card provisioning for HSBC UK within a 5-member API team — orchestrating multi-step customer journeys across customer-data, card and Visa services using reactive programming, asynchronous polling and robust failure handling.
Team hackathon project — built the Java backend: a layered DAO → service design over 8 MySQL tables covering subscription plans (daily to monthly), recurring deliveries and orders. Orders are generated on each delivery date instead of up front. 18 custom exceptions, injection-safe JDBC and 198 JUnit 5 + Mockito tests.
▶ Backend demo video ↗Extended an IIT Madras course project into a full-stack system with 25+ REST APIs, JWT auth and RBAC. Redis caching with TTL-based invalidation, Celery async tasks (Google Chat + SMTP reminders, PDF reports, CSV exports via 4-state job machine), 6 SQLAlchemy entities, full OpenAPI 3.0.3 spec, Vue 3 frontend with Chart.js dashboards.
Autonomous system that generates, executes and validates Python code for analytical questions over uploaded datasets using LangChain tool calling, multi-model fallback and subprocess-based sandboxed execution. Hardened with 4 failure layers — retry logic, quota-aware key rotation, graceful model degradation and generated-code error recovery.
RAG chatbot for IIT Madras' Tools in Data Science course — indexing 2,255+ content chunks into SQLite with embedding-based semantic search, multimodal (text + image) QA with cosine-similarity retrieval, deployed on Vercel.
AI-powered web app that instantly transforms raw text, notes or markdown into polished, custom-styled PowerPoint presentations. Uses LLMs to parse and structure content, then applies user-provided templates via python-pptx to generate professional decks ready to download.
1st place overall in my gym's 5K race, finished in 20 minutes. Endurance is my strong suit.
Training front lever, back lever, planche and handstand progressions, and sharing what I learn on Instagram.
Kerala, the Ellora & Ajanta caves and many more places. I share my trips on Instagram too.
Whether it's a backend challenge, a payment system, or an interesting GenAI project — I'm always open to connecting. Drop me an email or reach out on LinkedIn.
When I'm not coding, I'm training calisthenics. Watch on Instagram @samdoescal ↗