About
From pixels to infrastructure.
Full-Stack Engineer designing and delivering cloud-native products with Go, Vue, and AWS.
50,000+learners on the platform
99.9%uptime, held with Prometheus and Grafana
~8internal services behind one BFF
5m→30sp95 on the rescued dashboard
+12%retention and course completion
I'm Akash Aloysius James, a Full-Stack Engineer based in Milan. For the past several years I've built cloud-native products end-to-end — from the Vue.js and TypeScript interfaces people actually touch, to the Go services and AWS infrastructure that keep everything running underneath. At Lifeed, I work on an eLearning platform used by more than 50,000 people: I own the learner PWA's frontend, built the internal admin platform on my own from domain model through to interface, and rescued a client-facing analytics dashboard that had degraded to the point of timing out.
What I care about most are the parts of software users never notice: reliable systems, fast response times, and clean architecture that lets a team ship confidently without breaking things. My path here started with a Computer Science degree in Chennai and led to a Master's at Politecnico di Milano, which I worked through — building document-management and approval-workflow applications at MHS Global Impact alongside the degree. Whether I'm shaping API design in a planning session, tuning a frontend for accessibility and performance, or provisioning infrastructure as code with Terraform, I like owning problems from first idea to production. Always happy to connect with people who care about building thoughtful, well-crafted software.
Experience
October 2023
— Present
Milan, Italy
Full-Stack Engineer · Lifeed
Full-Stack Engineer on a Go + Vue/TypeScript eLearning platform serving 50,000+ learners — responsible for the learner PWA's frontend, and for the internal admin platform end-to-end, from domain model and interface design through to implementation.
Responsibilities
- Owned the learner-facing Vue/TypeScript PWA and the Go Backend-for-Frontend in front of it — the single trust boundary between the browser and the platform's internal services.
- Designed and maintained that BFF: aggregating a request across roughly eight internal services, verifying auth per request, and owning the session and token lifecycle so no credential ever reached the client.
- Led frontend development of the progressive web app (Vue, Quasar, GraphQL, Strapi CMS), including push notifications and reusable content.
- Independently built and operated an internal admin platform giving Customer Success self-serve tooling for previously manual operations.
- Worked across cloud infrastructure on AWS — EC2, RDS, ALB, and Docker.
- Owned observability, testing, and CI/CD across the stack (Prometheus, Grafana, Go testing, Jest, Vitest, GitLab).
- Partnered with Product, Design, Content, and Customer Success to turn requirements into shipped features.
- Managed third-party integrations — Intercom, Typeform, Wistia — and localization workflows via Ditto and Figma.
- Reviewed PRs, set coding standards, and contributed to architecture and roadmap decisions in technical planning.
Achievements
- Rescued a client-facing analytics dashboard that had stopped loading — main queries had drifted to roughly five minutes against a 60-second timeout. Query preloading, splitting the data fetch so rendering no longer waited on the slowest result, and a cache-warming job brought p95 load time to around 30 seconds.
- Held 99.9% uptime by introducing Prometheus and Grafana monitoring across the platform.
- Made a real Postgres the merge gate for anything touching the data layer, with test coverage aimed at the functionality where failure is costly rather than spread evenly to hit a percentage.
- Improved user retention and course completion by 12% through a rebuilt Intercom workflow with event tracking and email notifications.
- Built the learner-facing frontend to WCAG 2.1 AA standards through semantic HTML and dynamic ARIA attributes.
- Shipped an internal admin platform from scratch, automating operations that were previously manual.
- Took engineers off the critical path for copy changes — localized text moved from per-locale database rows, which required an engineer and a content creator working together, into Ditto connected to Figma frames, with a BFF-side script fetching and caching the published result.
- Migrated the platform's data hosting from the US to the EU, auditing every integration and workflow along the way — the audit is what surfaced the idle event-tracking capability that became the retention work.
Key Decisions
- Build an internal admin platform in-house rather than patch manual processes — traded upfront effort for lasting operational flexibility and self-serve tooling for non-engineers.
- Warm the analytics cache on a schedule rather than scale the database — the root cause was a subquery cache whose invalidation cascaded through every query layered above it, so the fix moved the expensive recomputation off the user's request instead of buying hardware to absorb it.
- Invest early in observability (Prometheus + Grafana) — made reliability measurable and turned 99.9% uptime into a defended target rather than a hope.
- Use the tool already paid for rather than buy another — the support platform already in the stack had event tracking and workflow automation sitting unused, which turned a procurement question into a configuration one.
- Adopt Quasar + Strapi CMS + GraphQL for the learning experience — decoupled content from code so non-technical teams could author dynamic material.
- Test against a real database rather than mocks where correctness lives in the schema — a mocked database validates your assumptions, and the bugs worth catching are the ones where those assumptions are wrong.
- Establish shared coding standards and drive API architecture in planning — prioritized long-term consistency and maintainability over short-term speed.
February 2021
— June 2022
Junior Full-Stack Developer · MHS Global Impact
Early-career full-stack role building document management and approval-workflow applications across web and mobile, using Angular and .NET Core on a MySQL backend.
Responsibilities
- Developed Angular applications and responsive interfaces for document management and approval workflows.
- Built and maintained .NET Core REST APIs powering those workflows.
- Optimized MySQL schemas and queries to improve backend performance.
- Created reusable UI components, data-visualization interfaces, animations, and responsive layouts.
- Maintained frontend performance, accessibility, and cross-browser compatibility.
- Delivered Android applications from the existing web codebase using Cordova.
- Contributed to AWS deployments and Jenkins CI/CD pipelines.
Achievements
- Shipped document management and approval-workflow applications end-to-end, across Angular frontends and .NET Core APIs.
- Improved backend performance by restructuring MySQL schemas and queries.
- Extended the product to Android without a separate native codebase by integrating Cordova.
- Built a library of reusable components and data-visualization interfaces that accelerated UI development.
- Broadened reach by improving accessibility and cross-browser compatibility across the frontend.
- Helped automate releases by contributing to Jenkins CI/CD on AWS.
Key Decisions
- Extend to mobile with Cordova instead of a native rebuild — reused the existing web codebase to ship Android quickly, with a single codebase to maintain.
- Optimize MySQL at the schema level, not just the query level — addressed performance at its source rather than patching slow queries one by one.
- Invest in a shared component library — traded a little upfront effort for UI consistency and faster feature delivery.
- Contribute to CI/CD automation — favored a repeatable Jenkins pipeline over manual, error-prone deployments.
Education
MSc, Computer Science & Engineering
Politecnico di Milano
BSc, Computer Science & Technology
SRM University
Skills & Capabilities
Frontend
TypeScriptVueQuasarReactNext.jsTailwind
Accessible product interfaces held to WCAG 2.1 AA — Vue and Quasar on the learning platform, React and RSC on this site.
Backend
GoNode.jsGraphQLSSE.NET Core
Streaming Go services with SSE fan-out to the browser, and GraphQL where content had to decouple from code.
Infra
AWSTerraformDockerCI/CDPrometheusGrafana
Keyless CI/CD to AWS over OIDC, the whole estate in Terraform, one box running Docker Compose and watched by Prometheus.
Data
PostgreSQLTimescaleDBRedisMySQL
Time-series at scale — Timescale hypertables for candles, Redis for caching and leader leases, and relational schemas tuned when queries drifted.