const engineer = {
  name: "Dhyey Patel",
  role: "Software Engineer",
  university: "Arizona State University",
  gpa: 3.78,
  focus: [
    "Cloud", "Distributed Systems",
    "AI/ML"
  ]
};

Cloud · Backend · AI/ML

Hello, I’m

Dhyey Patel

Software Engineer

San Francisco, CA

17.7K+decisions / day

99%precision

740decisions / hour

Get To Know More

About Me

Arizona State University

M.S. Computer Software Engineering
Jan 2025 – Dec 2026

GPA: 3.78 / 4.0Tempe, AZCloud Computing & AI

Gujarat Technological University

B.E. Information Technology Engineering
Oct 2020 – Jun 2024

GPA: 9.21 / 10.0Ahmedabad, IndiaML & Computer Vision

I’m a software engineer pursuing my Master’s at Arizona State University. I build reliable, cloud-native systems—from responsive full-stack applications to event-driven platforms that process production workloads at scale.

At Amazon Music, I built a decision platform processing 17.7K+ catalog decisions per day at 99% precision.

Experienced in AWS serverless architecture, Docker, Kubernetes, CI/CD, observability, and distributed systems.

Strong across Python, Java, C++, TypeScript, and JavaScript, with hands-on backend, frontend, and AI/ML experience.

Where I’ve Made an Impact

Experience

Amazon Music · Amazon.com, Inc.

Software Development Engineer Intern

San Francisco, CA

May 2026 – Aug 2026

Built a production Python decision platform processing 17.7K+ catalog decisions per day at 99% precision, reducing resolution time from days to 90 seconds.
Engineered automated retries, failure isolation, audit logging, and monitoring, achieving zero failures across 142K+ production decisions.
Architected AWS Lambda, Step Functions, SQS, and DynamoDB workflows sustaining 740 decisions per hour without downstream throttling.
Helped design a scaling path from 17.7K to 88K+ decisions per day.

Xmatiq Technologies · Techforever Solutions LLP

Full-Stack Developer Intern

Ahmedabad, Gujarat, India

Jan 2024 – Apr 2024

Developed Node.js, Express.js, and MongoDB services and REST APIs for ERP systems, improving data-processing efficiency by 80%.
Built a PHP and MySQL web application with validation, error handling, and third-party payment integration.

Ingenious Technologies

ReactJS Intern

Vadodara, Gujarat, India

Jul 2023 – Aug 2023

Implemented JWT-secured REST APIs supporting 200+ requests per minute with 99% uptime.
Added lazy loading and optimized state management, reducing page-load time by 30%.

My Technical Toolkit

Skills

Languages

PythonGoOCamlC++JavaSQLTypeScriptJavaScript

Systems & Backend

LinuxFastAPISpring BootNode.jsRESTful APIsCeleryPydanticOpenAPI

Cloud & Distributed Systems

AWSDockerKubernetesKueuecontroller-runtimeLambdaSQSStep FunctionsDynamoDBS3

Databases & Caching

PostgreSQLMySQLMongoDBRedisSQLiteSQLAlchemyAlembic

Frontend

ReactNext.jsTailwind CSSHTML5CSS3

DevOps & Observability

GitGitHub ActionsJenkinsCI/CDDocker ComposekindKustomizePrometheusGrafana

AI & Machine Learning

PyTorchTensorFlowTransformersAWS BedrockScikit-learnModel Context Protocol (MCP)LLM Tool UseLLM Evaluation

Testing & Verification

pytestPlaywrightVitestJestMockitoPostmanDifferential TestingFault Injection

Security & Reliability

SandboxingRBACSHA-256 Hash ChainsIdempotencyOptimistic LockingWrite-Ahead LoggingCrash Recovery

Selected Engineering Work

Projects

FastAPI · PostgreSQL · MCP

Release candidate · Backend & AI lead

Continuum AI

Decision memory for AI assistants

Owned the backend for a two-person project that captures decisions and their reasons, then detects when changing assumptions break a plan. Built evidence-grounded extraction, dependency-graph traversal, versioned revisions, and eight MCP tools sharing the REST service layer.

40/40labeled detection-engine cases passed

Python · FastAPI · Next.js · Docker

Demo · Backend, security & infrastructure lead

ProofLayer

Independent verification for AI coding agents

Built run orchestration, sandbox and verifier drivers, and a SHA-256 hash-chained evidence engine for a two-person platform. A separate verifier re-runs pinned checks, and deterministic rules classify agent claims as verified, failed, or inconclusive, with downloadable evidence certificates.

142automated backend & frontend tests

Go · Kubernetes · Kueue

Implemented · Locally verified

KubeFleet

Batch admission & lifecycle operator

Built a Go operator with immutable BatchRun APIs, UID-safe Job ownership, prerequisite admission checks, and durable cancellation and deadline cleanup. Added a tenant CLI, restricted worker Pods, and observability; exercised process crashes and leader failover on a local kind cluster.

180successful local smoke batch runs

View source

OCaml · Python · WAL · Docker

v0.1.0 · Local CLI workbench

ReplayLab

Matching engine & crash-recovery workbench

Built an OCaml price/time-priority matching engine checked against an independent Python reference. Implemented synchronous journaling, snapshots, retry-safe recovery, and nine controlled crash hooks. Failure-reduction tooling shrank a seeded FIFO defect from 500 attempts to three commands.

100Kattempts in the differential test corpus

View source

PyTorch · FAISS · FastAPI · Kafka

In progress · Product & experiment design

InspireRank

Multi-interest visual recommendations

Designing a recommendation benchmark and interactive feed that combine multiple user interests with recent activity. The planned system compares retrieval and ranking models with chronological evaluation, session adaptation, and diversity reranking. Implementation and measured results are pending.

PyTorch · WavLM · Scikit-learn

〽️

Aug 2025 – Present

Speaker-Independent Speech Emotion Recognition

Built an eight-class, speaker-independent speech inference pipeline with pretrained WavLM representations, statistical pooling, PCA, and downstream classification, achieving 82% test accuracy.

82%test accuracy

AWS EKS · Terraform · Helm

☁️

Aug 2025 – Dec 2025

Cloud-Native Microservices Platform

Deployed containerized Python services to AWS EKS with Docker, Kubernetes, Terraform, and Helm. Added CI/CD plus Prometheus and Grafana monitoring for repeatable deployments and operational visibility.

React · FastAPI · PostgreSQL

🎙️

Feb 2025 – Jun 2025

NextMeet-AI Meeting Companion

Built an AI meeting companion for transcription, summaries, and action items. Celery and Redis decouple long-running AI workloads while the frontend presents responsive, real-time task progress.

React · TypeScript · Cloudflare

✦

Sep 2025 – Oct 2025

Nebula Spotlight

Built an AI-powered Chrome extension that automates browser interactions from natural-language commands using a universal DOM element-resolution framework. Added secure Cloudflare Worker APIs, client-side PII masking, caching, and prompt orchestration.

Lambda · SQS · DynamoDB

⚡

Jan 2026 – Mar 2026

Event-Driven Order Management System

Architected independently scalable order, inventory, and analytics services. Implemented idempotent consumers, queue-based decoupling, authentication, and failure handling for reliable distributed workflows.

Explore my GitHub ↗

Get in Touch

Contact Me

Have a role, project, or interesting engineering problem in mind? Let’s talk.

San Francisco, CA

+1 623 800 2657

dhyey0201@gmail.com