Backend Engineer

Ian Legaria
Morales

Software Engineer III · Currently at Guru

Seven years building backend systems across enterprise, SaaS, and startup contexts — from greenfield microservices to legacy modernization, complex API integrations to large-scale data workflows. Consistent focus on correctness, reliability, and maintainability regardless of the stack.

Case Studies GitHub ↗ Resume ↓
Technical Profile

What I work with

Java
Primary language · 7+ years
Spring Boot
Microservices · REST · GraphQL
JPA / Hibernate
Criteria API · Metamodel type-safe
PostgreSQL
Schema design · query optimization
Elasticsearch
Index design · analyzers · queries
Redis
Caching · rate limiting
Cloudant NoSQL
IBM Cloud · document store · IBM
AWS
SQS (FIFO) · async messaging · cloud infra
Google Pub/Sub
Event-driven workflows
OAuth 2.0
Auth code · client credentials
GraphQL
API design · schema-first
Go (Golang)
REST APIs · feature flags · IBM
TypeScript · Node
Integration platform · API clients
GCP
GCS · Pub/Sub · GKE · Cloud Run
Kubernetes
GKE · container orchestration
Docker · Terraform
Containerization · IaC
Stripe
Payment processing · webhooks
Grafana · Prometheus
Observability · on-call
Liquibase
Schema migrations · versioning
Jenkins · Travis CI
CI/CD pipelines
Selected Work

Case Studies

138
Pull requests@ Guru
662
Commits shipped@ Guru
10+
APIs integrated
7yr
Backend experience
01 / 05 Guru Technologies
Zero-Downtime Data Migration
Restructuring a live content model in production — dual-write · idempotent backfill · schema cleanup
Problem

A core content entity was storing structured data as raw JSON in a single column. As the product evolved, that model needed to be broken out into a normalized relational structure — but the table was live, under constant write pressure, and a big-bang migration wasn't an option.

Approach

Implemented the dual-write pattern: all new writes went to both the old and new schemas simultaneously. A background admin job backfilled historical records in batches, with duplicate detection on each pass to make reruns safe. Once data was verified, reads were cut over, legacy columns dropped, and old tables removed.

Outcome

Full migration completed with zero downtime and no data loss. The new schema is normalized, independently queryable, and decoupled from the parent entity — making future features significantly cheaper to build.

JavaPostgreSQLJPADual-Write PatternBatch ProcessingIdempotency
02 / 05 Guru Technologies
Granular Authorization System
RBAC with custom roles and atomic permissions across multiple resource types
Problem

The existing authorization model only supported a fixed set of global roles. Enterprise customers needed workspace-level customization — the ability to define roles with precise, composable permissions spanning different resource types.

Approach

Implemented a full RBAC system: a CRUD API for custom role definitions, an atomic permission matrix covering granular actions per resource type, and per-resource permission endpoints. Access validation logic extracted into a dedicated service — consolidating five distinct access paths into a single testable unit. Rollout gated behind a feature flag. Orphaned access records cleaned up automatically on role deletion.

Outcome

A complete authorization layer covering the full role lifecycle, permission enforcement across all resource types, and automated access cleanup — deployed progressively with no breaking changes to existing behavior.

JavaRBACREST API DesignFeature FlagsAccess ControlPostgreSQL
03 / 05 Cision
Enterprise Search Infrastructure
Designing an Elasticsearch layer from scratch for a greenfield content platform — index design · analyzers · query architecture
Problem

A new microservices platform serving journalists, content creators, and enterprise users had no search infrastructure. The content model was heterogeneous — multiple content types with different field structures — and search had to be built from the ground up to match the platform's query requirements from day one.

Approach

Designed and implemented the Elasticsearch integration from scratch: defined index mappings aligned with the content model, configured analyzers and tokenizers for full-text search behavior (stemming, partial match, language handling), and built the query architecture for enterprise content retrieval. Implemented as a dedicated service integrated with the Spring Boot backend.

Outcome

Full search infrastructure operational from platform launch. The index design supported the evolving content model without requiring structural changes as new features were added. Search became a stable foundational layer that other services could depend on.

JavaSpring BootElasticsearchIndex DesignAnalyzersMicroservicesGCP
04 / 05 IBM
Feature Flags System in Go
Building a Golang-based feature flag service for incremental rollouts on an enterprise contracts platform
Problem

An internal IBM contracts platform had no mechanism to roll out new functionality incrementally. Every release carried the full risk of all its changes — features couldn't be toggled per user or context without a redeployment, making controlled releases on enterprise-critical workflows impractical.

Approach

Collaborated with a colleague to design and implement a Golang-based feature flags service. Built the flag evaluation logic, the management API, and the integration layer with the existing contracts backend. The system supported flag targeting by user and context attributes, enabling controlled rollouts and gated releases across enterprise workflows.

Outcome

Feature rollouts decoupled from deployments. New functionality could be exposed to specific users before wider release, reducing risk on a platform handling enterprise contracts. Test coverage of the contracts backend also increased from 20% to 80% over this period.

GoREST APIFeature FlagsCloudant NoSQLDockerIBM CloudTravis CI
05 / 05 Guru Technologies
Multi-Platform Integration Layer
10+ external API integrations with OAuth 2.0, rate-limit retry, and incremental sync
Problem

The platform needed to sync content from a growing list of third-party SaaS tools — Salesforce, Jira, Confluence, Coda, Smartsheet, and others — each with different auth models, pagination strategies, rate limits, and data hierarchies.

Approach

Built each integration on a unified architecture: OAuth 2.0 (auth code and client credentials), API key, Basic, and Bearer auth — consistent across connectors. Pagination handled both cursor-based and page-number strategies. Rate-limit retry with exponential backoff where required. Sync: incremental every 4h, full weekly, initial on connection. For one connector, implemented a dual-path architecture detecting tenant API capabilities at runtime. Also built an internal scaffolding tool that reduced the effort to add a new connector from days to hours.

Outcome

Consistent integration coverage across 10+ platforms. The capability-based routing pattern became reusable. The scaffolding tool enforced structural consistency and cut onboarding time for new integrations.

TypeScriptOAuth 2.0Rate LimitingIncremental SyncDual-Path Architecture
Career

Experience

Jul 2024 — Present · Guru Remote · Philadelphia
Software Engineer III (Backend)
  • Implemented a zero-downtime migration of a core content model using the dual-write pattern — batched backfills, idempotent reruns, schema cleanup post-cutover.
  • Built a custom RBAC system: role CRUD API, atomic permission matrix across multiple resource types, and orphaned access record cleanup gated behind a feature flag.
  • Built and maintained 10+ integrations with external SaaS APIs (Salesforce, Jira, Confluence, GitHub, Coda) — handling OAuth 2.0, pagination, rate-limit retry, and incremental sync.
  • Built an internal tool powered by AI agents (Claude Skills, LangGraph) to automate the development, deployment, and debugging of integrations — cutting average delivery time by over 60%.
  • Adopted AI-assisted development workflows using Claude, ChatGPT, and Cursor throughout the engineering cycle — from design and implementation to debugging and code review — as standard tooling for daily backend work.
  • Migrated multiple services from deprecated Hibernate Criteria API to type-safe JPA Criteria with Metamodel attributes.
JavaSpring BootJPA / Hibernate PostgreSQLAWS SQSElasticsearch TypeScriptOAuth 2.0Liquibase
May 2022 — Jul 2024 · Cision Remote · Mexico City
Software Engineer II
  • Joined a greenfield microservices platform built from scratch for content creators, journalists, and enterprise users — ~30 engineers across multiple service teams.
  • Co-designed one of the platform's first GraphQL services: a media pipeline that accepted image uploads, generated multiple resized variants, and stored results in Google Cloud Storage.
  • Built the Elasticsearch search layer from scratch: index mappings, analyzers, tokenizers, and query structure for enterprise content search.
  • Integrated Stripe for payment processing and wired Google Cloud Pub/Sub for the async subscription update flow triggered on purchase events.
  • Deployed on GKE via GitHub Actions and Terraform; monitored with Grafana and Prometheus. Participated in on-call rotation.
JavaSpring BootGraphQL PostgreSQLElasticsearchPub/Sub StripeKubernetesTerraform
Sep 2019 — May 2022 · IBM Remote · Guadalajara
Full Stack Developer
  • Worked on two internal IBM platforms across separate teams — both modernization projects involving legacy systems and enterprise-scale integrations.
  • On the first project: built and maintained backend services in Go and REST APIs for an internal contracts system; increased test coverage from 20% to 80%; collaborated on a Golang feature flags system.
  • On the second project: contributed to migrating a legacy Java 7 / Servlets app hosted on IBM WebSphere to IBM's hybrid cloud — dockerized the application, introduced Jenkins for CI, and helped start a new React frontend.
GoJavaReact Cloudant NoSQLPostgreSQL DockerKubernetes JenkinsIBM Cloud
Jul 2018 — Sep 2019 · SmartSoft America Tlaxcala, Mexico
Java Developer
  • First professional role — worked across multiple client projects including a government cadastral system, an events and conferences platform, and a cargo logistics app for container management.
  • Built backend systems using Java, JSF, and REST APIs; deployed on GlassFish application server.
  • Developed Android mobile applications in Kotlin, integrated with backend services.
  • Designed and maintained MySQL and PostgreSQL databases across multiple projects.
JavaJSFKotlin AndroidMySQL PostgreSQLGlassFish
Independent Engineering

Personal Projects

Featured Project

Mercury

Backend Systems Case Study

Self-hosted RAG backend for document ingestion, semantic search, grounded Q&A, quiz generation, snippets, and spaced-repetition flashcards.

Backend focus

A self-hosted knowledge engine designed to organize and study personal documents using semantic search. Created out of a real need, and used as a platform to explore backend system design concepts like async pipelines, caching strategies, and LLM integrations.

Highlights
  • Spring Boot 3.5, Java 21, PostgreSQL + pgvector
  • RabbitMQ ingestion pipeline with transactional outbox
  • Redis search caching and Testcontainers integration tests
  • GitHub Actions backend CI and protected main branch
JavaSpring BootPostgreSQLpgvectorRabbitMQRedisTestcontainersGitHub Actions
Repository ↗ Architecture ↗ ADRs ↗ CI ↗
Background

Education & Publications

Aug 2014 — Jul 2018 · UAT Tlaxcala, Mexico
B.S. Computer Engineering
  • Universidad Autónoma de Tlaxcala — Computer Engineering program.
Academic Publication · IZTATL Magazine ISSN: 2007-9958
Risk Management in the Development of Software Projects
  • Published research article on risk identification and mitigation strategies in software project development, featured in IZTATL academic magazine.
Get in touch

Contact

Emailil.legaria@gmail.com
GitHubgithub.com/iLegaria
LinkedInlinkedin.com/in/ian-legaria-morales
ResumeDownload PDF ↓