Available for opportunities

Aman
Goyal

I build the infrastructure
AI systems run on.

Hire me, and I promise to only break things in staging... most of the time.

View My Work Get In Touch
Aman Goyal
3+ Startups Helped
$6K+ Saved Per Year
0 Downtime Deploys
100% Infra as Code
What I Do

Stack. Ship.
Scale.

01

Cloud Infrastructure

AWSECS FargateTerraformRDSCloudFrontCI/CD

Production-grade AWS environments — VPC, ALB, ECS/Fargate, RDS, CloudFront, WAF. Blue-green CI/CD with automated rollbacks, full IaC via Terraform, and zero-downtime deploys as the baseline.

02

Backend Systems

Node.jsRedisBullMQInngestPostgreSQL

Fault-tolerant async job architectures for AI media pipelines, credit systems, and real-time orchestration. Built to handle concurrent workloads under load without degradation.

03

AI Systems & Agents

OpenAIGeminiMistralEmbeddingsAgentic PipelinesBYOK

LLM-powered products built around how agents interact with software — browser-native agents that inject into and operate within live interfaces, semantic retrieval pipelines, agentic execution flows, and BYOK abstraction layers with real user traction.

04

Product Engineering

TypeScriptNext.jsChrome ExtensionsClerkPayments

Full-stack products shipped end-to-end: SaaS platforms, Chrome extensions, auth, billing, and usage metering — from first commit to live users, without a team behind me.

Work That Shipped

Real problems.
Real outcomes.

01 AI Chrome Extension
10+ Waitlist
Signups

Lazyapply

lazyapply.xyz
Lazyapply extension UI inside Twitter/X reply composer
Problem

Founders doing outbound distribution on X and Reddit were forced to manually copy post context into a separate AI chat, breaking their workflow and producing generic, low-signal replies at scale.

What I Built

Chrome extension with content scripts that programmatically inject into live software interfaces (X and Reddit), extract structured state from the DOM, augment it with stored context, and surface AI output inside the native UI — all without modifying the host application.

Result

Organically growing waitlist of 10+ signups ahead of Chrome Web Store approval. Strong early traffic signals validate demand. Freemium credit model with usage limits and billing logic built and ready to scale on launch.

Chrome ExtensionContent ScriptsTypeScriptPrompt EngineeringAuth/BillingFreemium
02 Open-Source AI Tool
50 Active
Users

BookmarkBrain

chromewebstore.google.com GitHub
BookmarkBrain semantic search interface
Problem

Saved posts and bookmarks on X and Reddit accumulate as an unsearchable archive — no retrieval, no synthesis, and no practical way to resurface relevant content weeks later.

What I Built

Privacy-first Chrome extension with local-first storage architecture, semantic embeddings over saved content, natural language Q&A returning source-cited answers, and a BYOK abstraction layer supporting OpenAI, OpenRouter, and Gemini.

Result

50 active users with zero backend dependency for personal data — all storage and retrieval runs in-browser. Fully open-source and publicly shipped. Multi-provider AI support makes it model-agnostic and future-proof.

Chrome ExtensionSemantic SearchEmbeddingsVector RetrievalBYOKOpen Source
03 AI Video SaaS
0 Downtime
on Deploys

Vubo

vubo.ai
Problem

Needed production-grade, multi-tenant AI video SaaS infrastructure built from the ground up — no existing playbook, aggressive timelines.

What I Built

AWS ECS Fargate + Redis/BullMQ/Inngest pipeline for AI media processing, subscription-gated credit system, CI/CD via GitHub Actions, OpenTelemetry tracing across every pipeline step.

Result

Zero-downtime deployments with automated blue-green rollouts. Full observability across AI generation workflows. System handles concurrent workloads without degradation.

AWS ECSNext.jsRedisInngestBullMQGitHub Actions
04 Startup Cost Reduction
$6K+ Saved
Per Year

Self-Hosted Stack

Problem

Early-stage startups burning runway on SaaS analytics, live chat, uptime monitoring, and session replay tools — spending before finding product-market fit.

What I Built

Replaced paid SaaS with self-hosted Umami (analytics), Chatwoot (support), Uptime Kuma (monitoring), OpenReplay (session replay) on a single hardened VPS with full HTTPS/TLS.

Result

Over $6,000/year recaptured with zero functional compromise. Least-privilege access, VPN-restricted internals, environment isolation — more secure than the SaaS it replaced.

DockerNginxVPSUmamiChatwootOpenReplay
05 AI Video Analysis Platform
2 AI Models
Orchestrated

ClipDecode

clipdecode.com
Problem

Extracting meaningful insights from video content required manual review — slow, unscalable, and completely unusable for teams processing large volumes of clips.

What I Built

4-stage AI pipeline (download → transcription → analysis → synthesis) using Google Gemini + Mistral APIs, orchestrated via Convex scheduler with fault-tolerant retry logic and real-time state updates.

Result

Fully automated video-to-insights pipeline. Credit-based usage with tier access, idempotent Dodo Payments webhooks for zero-loss billing, and Clerk auth — shipped as a production SaaS.

Next.js 15ConvexGeminiMistralTypeScriptDodo Payments
06 Infrastructure as Code
100% Infrastructure
as Code

Cloud Infra Automation

Problem

Manual provisioning was error-prone, non-reproducible, and created dangerous drift between staging and production environments.

What I Built

Modular Terraform for VPC, ALB, ECS/Fargate, RDS, CloudFront, WAF, ElastiCache. Blue-green deployments via CodeDeploy integrated with GitHub Actions and ECR workflows.

Result

Fully reproducible, secure-by-default infra with least-privilege IAM, WAF rules, and SSM secrets management. New environments spin up in minutes. Zero configuration drift.

TerraformAWSCodeDeployCloudWatchWAFRDS Proxy
Experience

Where I've
shipped.

Founding Engineer / DevOps & Backend Full-time
Vubo ↗
Feb 2025 — Present
  • Built and maintained production infrastructure for an AI-powered video SaaS using AWS (ECS Fargate, S3, CloudFront, RDS, Lambda) and Docker-based deployments.
  • Designed scalable background job architecture using Redis + BullMQ/Inngest for media processing, AI workflows, and async task orchestration.
  • Implemented subscription-based credit system with atomic usage control, rate limiting, and workspace-level storage enforcement.
  • Optimized CI/CD pipelines using GitHub Actions with build caching, environment automation, and automated CloudFront cache invalidation.
  • Integrated observability with OpenTelemetry and PostHog to monitor performance and user behavior across AI generation pipelines.
  • Reduced infrastructure costs through architecture optimization, resource right-sizing, and efficient cloud asset delivery strategies.
Software Engineer — Freelance / DevOps Freelance
Independent Projects & Startup Consulting
Jun 2024 — Present
  • Designed and deployed cloud infrastructure for early-stage startups using AWS, Docker, reverse proxies, and automated CI/CD workflows.
  • Reduced operational costs by replacing $500+/month SaaS tools with self-hosted alternatives (Umami, Chatwoot, Uptime Kuma, OpenReplay) on a single VPS.
  • Implemented secure production environments with HTTPS/TLS, least-privilege access, environment isolation, and VPN-restricted internal services.
  • Built backend services and automation workflows for analytics, media processing, and scheduled job execution with high reliability.
  • Provided end-to-end DevOps setup including infrastructure provisioning, monitoring, logging, and deployment automation for multiple projects.
Tech Stack

Tools I ship
production with.

Languages
TypeScriptJavaScriptPythonGolangSQLShell Scripting
Cloud & Infra
AWS ECS FargateEC2S3CloudFrontRDSLambdaIAMCloudWatchWAFDockerTerraformGitHub Actions
Backend
Node.jsNext.jsExpress.jsRedisBullMQInngestConvexPostgreSQLDrizzle ORMREST APIs
Observability
OpenTelemetryPostHogCloudWatchLoggingMetricsPerformance Optimization
Contact

Let's build
something fast.

Whether you're scaling infra, cutting costs, or shipping a new product — I've done it before and I'll get it right.

Download Resume amangoyel0@gmail.com
LinkedIn GitHub Portfolio