currently building: TODO(saketh): a RAG service on FastAPI + Cloud Run (sample text)
// sample content, pending review
Fig. 1 · Reference sketch of a service on Cloud Run
01 · About
Sheet 01 / 05
TODO(saketh): 3-4 sentence bio. Sample: I'm a backend-leaning software engineer who likes turning fuzzy requirements into well-tested services. Most of my work is Java/Spring Boot and Python/FastAPI running on Google Cloud, with enough React and TypeScript to ship the front end too. Lately I've been applying that foundation to AI/ML systems: retrieval pipelines, evaluation, and the infrastructure that keeps models useful in production.
02 · Skills
Sheet 02 / 05
Backend
Java
Spring Boot
Python
FastAPI
REST
gRPC
JUnit
pytest
08 items
08 items
Cloud / Infra
GCP
Cloud Run
Docker
Terraform
Cloud Build
Pub/Sub
06 items
06 items
Data
PostgreSQL
BigQuery
Redis
SQL
04 items
04 items
AI / ML
Embeddings
RAG
pgvector
LLM APIs
Evaluation
05 items
05 items
Frontend
TypeScript
React
Next.js
Tailwind CSS
04 items
04 items
Tooling
Git
GitHub Actions
Linux
Bash
04 items
04 items
03 · Experience
Sheet 03 / 05
REV C
Software Engineer, Backend
TODO(saketh): start date – Present
TODO(saketh): Employer A
TODO(saketh): one line on the team and product (sample).
Cut p95 latency of a Spring Boot order API by TODO(saketh): X% by adding Redis caching and fixing N+1 queries
Migrated TODO(saketh): N services from VMs to Cloud Run with Cloud Build CI/CD, reducing deploy time to TODO(saketh): X min
Built a FastAPI service for document search with pgvector embeddings, serving TODO(saketh): X requests/day
REV B
Software Engineer
TODO(saketh): start date – TODO(saketh): end date
TODO(saketh): Employer B
Designed REST APIs in Java/Spring for TODO(saketh): domain, covered by JUnit and Testcontainers integration tests
Built Pub/Sub to BigQuery event pipelines processing TODO(saketh): X events/day
Shipped React + TypeScript internal tools that TODO(saketh): outcome
REV A
Software Engineering Intern
TODO(saketh): start date – TODO(saketh): end date
TODO(saketh): Employer C
Automated TODO(saketh): workflow with Python scripts, saving TODO(saketh): X hours/week
This portfolio renders from a single resume.json. The goal: a new resume is parsed, validated with zod and lands as a reviewed pull request before Cloud Build deploys it to Cloud Run. Validation and rendering work today; the pull request and deploy stages are in progress.
resume.pdfinbox
Parsezod
resume.jsonvalidated
PR reviewhuman gate
Cloud Rundeploy
Fig. 2 · Resume-to-site pipeline (PR review and deploy in progress)