Open to senior / staff roles

YashVaidya

Building AI / LLM Systems

I build production AI systems, multi-tenant SaaS platforms, and the self-hosted infrastructure that runs them — from RAG pipelines to event-driven backbones to bare-metal Proxmox clusters.

Get in touch Download résumé
Richmond, VA
FastAPI✦Next.js✦Kafka / Redpanda✦PostgreSQL✦Celery✦React✦RAG / LLMs✦Docker✦Proxmox✦RabbitMQ✦Redis✦Go✦pgvector✦AWS✦Traefik✦MQTT✦Tailscale✦BullMQ✦TimescaleDB✦FastAPI✦Next.js✦Kafka / Redpanda✦PostgreSQL✦Celery✦React✦RAG / LLMs✦Docker✦Proxmox✦RabbitMQ✦Redis✦Go✦pgvector✦AWS✦Traefik✦MQTT✦Tailscale✦BullMQ✦TimescaleDB✦

About

I build the platform and the system that runs it.

Full-Stack & Platform Engineer with 3+ years building production AI systems, multi-tenant SaaS platforms, and self-hosted infrastructure. I've shipped LLM-powered document intelligence and RAG pipelines for enterprise clients, architected event-driven B2B2C platforms on FastAPI, Next.js, Redpanda and PostgreSQL, and published open-source DevOps tooling for credentials management and zero-trust SSH access.

I'm comfortable across the whole stack — from React frontends to async Python services, Kafka/Redpanda event backbones and Celery workers, down to Kubernetes and a bare-metal Proxmox homelab. AWS Certified. I tend to own problems end to end: the customer conversation, the architecture, the code, and the 2am production incident.

RAG / LLM systemsEvent-driven architectureMulti-tenant SaaSRBAC + ABACFastAPI · Next.jsKafka / RedpandaCelery · BullMQPostgreSQL · pgvectorProxmox homelabAWS serverless
0+
Years shipping production
0+
Platforms built in 2026
0
Open-source tools published
AWS
Certified Cloud Practitioner

Education

  • M.S. · Engineering Sciences — Robotics
    University at Buffalo (SUNY)
    2023
  • B.S. · Computer Engineering
    K. C. College of Engineering, Mumbai University
    2022
  • Diploma · Computer Engineering
    Maharashtra State Board of Technical Education
    2019

Certifications

  • AWS Certified Cloud Practitioner
    Amazon Web Services · 2024
  • AWS Certified Solutions Architect
    Amazon Web Services · In progress

Experience

Where I've had impact

Jan 2024 — Present

ITHENA

Full-Stack & DevOps Engineer

Richmond, VA

AI · Cloud · Frontend · Backend

Own end-to-end delivery across AI/ML, cloud, frontend, and backend for enterprise and manufacturing clients — from the customer call to production.

  • Architected an LLM document-intelligence platform — RAG pipelines over 10K+ enterprise documents with vector embeddings and sub-2s semantic search — for clients incl. Case Engineering, Acuvi & Ignition.
  • Owned the company's on-prem build-out (architecture → procurement → cutover) and hosted in-house LLMs on it so manufacturing clients' data never leaves the environment.
  • Shipped a service-management/ticketing platform (React + Express) with AWS SES notifications and Google Pub/Sub event-driven workflows.
  • Built a B2B e-commerce platform with Three.js 3D product visualization (React, Express, Sequelize/MySQL on EC2).
  • Designed industrial IoT analytics (ThingsBoard, FactoryTalk, MQTT) with ML anomaly detection feeding predictive dashboards.
  • Built AWS automation (Lambda, Step Functions, DynamoDB) that cut idle compute cost, plus CI/CD + Docker pipelines across services — and informally mentor 4–5 engineers.
Next.jsReactPythonLangChainAWSDockerThree.jsPostgreSQL
Feb 2023 — Dec 2023

University at Buffalo (SUNY)

Research Assistant — Computer Vision

Buffalo, NY

Computer Vision

Computer-vision research and prototyping for object detection, segmentation, and facial recognition.

  • Built real-time vision systems (YOLO, Detectron2, OpenCV) for detection and segmentation across image/video streams.
  • Developed facial-recognition pipelines combining deep learning with classical CV, then validated and optimized for accuracy.
PythonOpenCVYOLODetectron2
Jun 2021 — Aug 2021

GodJN Solutions

Web Developer

Mumbai, India

Frontend · SEO

Frontend development and UI/UX modernization.

  • Led UI/UX modernization and SEO — 30% growth in organic traffic and a 25% lift in user engagement.
  • Built an Android app visualizing live accelerometer/gyroscope sensor streams.
JavaScriptHTML/CSSAndroid

Selected Work

Platforms I've shipped

Event-driven, multi-tenant systems built end to end — stacks read straight from each repo, not embellished.

01 Flagship
Manufacturing· 2026
0
isolated module databases
Private

Servix

Lead Engineer

B2B2C field-service management platform for industrial OEMs.

Event-DrivenMulti-tenantRBAC + ABAC
  • Single-tenant-per-OEM architecture with 13 isolated module databases serving post-sale service workflows.
  • Event-driven backbone on Redpanda (Kafka) with a transactional outbox — sub-second cross-module propagation of service-ticket lifecycles.
  • Hybrid RBAC + ABAC permissions with module:resource:action:scope granularity across multi-OEM deployments.
  • Async FastAPI + Celery workers, Alembic migrations, MinIO storage, and Grafana/Loki observability.
FastAPINext.js 14SQLAlchemyRedpandaCeleryPostgreSQLRedisMinIODocker
02
Full-Stack· 2026
0
FastAPI microservices
Private

B2BCom

Multi-tenant, event-driven B2B2C spare-parts commerce platform.

MicroservicesEvent-DrivenMulti-tenant
  • 11 FastAPI microservices (IAM, catalog, inventory, cart, order, payment, search, notification, audit…) behind a Traefik gateway.
  • Event-driven via Kafka + RabbitMQ with shared event/messaging libraries; per-OEM tenancy scaling to millions of SKUs.
  • Three Next.js apps — Super-Admin, OEM-Admin, and Storefront — on a shared design system, with per-tenant ERP/CRM ETL.
FastAPINext.jsKafkaRabbitMQTraefikPostgreSQLRedisMinIO
03
Security· 2026
0+
credential types
Open source

Vaulx

Self-hosted, multi-tenant credentials & secrets manager.

Open SourceZero-TrustMulti-tenant
  • 48+ credential types with AES-256-GCM encryption and per-user RSA key wrapping.
  • Multi-tenant Postgres row-level security; six-role RBAC + per-folder grants; SSO via SAML 2.0 & OIDC; WebAuthn passkeys.
  • BullMQ background jobs on Redis, full audit logging, and access-request approval workflows.
ReactNode / ExpressPrismaPostgreSQLBullMQSAML / OIDCWebAuthnMinIO
04
DevOps· 2026
0
static SSH keys
Open source

Shellius

Zero-trust SSH & RDP access with short-lived certificates.

Open SourceZero-TrustFleet Access
  • Per-org Ed25519 SSH Certificate Authority issuing short-lived signed certs — eliminates static authorized_keys.
  • Policy engine + manager approval workflow; browser SSH (xterm.js / WebSocket) and RDP via Apache Guacamole.
  • Companion Go (Bubble Tea) TUI client, OIDC login, and Prometheus metrics.
ReactNode / ExpressPrismaPostgreSQLssh2GuacamoleGo (Bubble Tea)BullMQ
05
AI / ML· 2024–25
0K+
documents indexed
Client engagement

AI Document Intelligence

LLM document-intelligence & RAG for enterprise (ITHENA clients).

RAGLLMEnterprise
  • RAG pipelines over 10K+ documents (PDF / text / structured) with vector embeddings and sub-2s semantic search.
  • Document-grounded Q&A chat with file upload and usage dashboards (Apache ECharts).
  • Tuned for low-latency, secure, production use across multiple enterprise clients.
Next.jsPythonLangChainOpenAIVector DBRAG
06
AI / ML· 2026
0
DB engines supported
Private

Ledgrr

AI-native, self-hostable personal-finance platform.

AI CategorizationSelf-hosted
  • Async FastAPI with a multi-dialect query runner (PostgreSQL / MySQL / SQL Server / Mongo).
  • Plaid bank sync + continuous AI transaction categorization with a deterministic NL-compiled rules engine.
  • arq job queue, OAuth + MFA, MinIO/Azure storage; React + TanStack Query/Table UI.
FastAPIReactSQLAlchemyarqRedisPlaidPostgreSQLMinIO
07
AI / ML· 2026
Open source

Docx

Document management with offline-first PWA and local LLMs.

Open SourceOffline PWALocal LLM
  • pgvector semantic search; local LLMs via Ollama plus Anthropic/OpenAI; Gotenberg document conversion.
  • Offline-first PWA (IndexedDB / Dexie) with a BlockNote block editor and arq background workers.
  • PDF rendering, markdown + syntax highlighting, and per-user storage on MinIO.
FastAPIReactpgvectorOllamaAnthropic / OpenAIarqDexieMinIO
08
DevOps· 2026
Open source

iAIOM

Open-source infrastructure & IoT monitoring SaaS.

Open SourceTime-SeriesEvent-Driven
  • MQTT ingestion (Mosquitto / aiomqtt) into TimescaleDB time-series storage.
  • Celery workers + beat for scheduled checks; Google SSO and rate limiting.
  • Drag-and-drop dashboards (react-grid-layout, Recharts) over a FastAPI + React stack.
FastAPIReactTimescaleDBMQTTCeleryRedis
09
IoT· 2024–25
Client engagement

Industrial IoT Analytics

Real-time industrial telemetry & predictive analytics (ITHENA).

TelemetryManufacturingML
  • Telemetry via ThingsBoard, FactoryTalk DataMosaix/Optix and MQTT; high-frequency time-series normalization.
  • ML anomaly detection and early fault identification feeding predictive dashboards.
ThingsBoardFactoryTalkMQTTPythonML
10
Full-Stack· 2026
Private

Nexride

Mobility platform concept with a live map interface.

In ProgressMaps
  • React + Vite experience with MapLibre GL interactive maps and framer-motion.
  • Early-stage foundation for a ride / mobility product.
ReactViteMapLibre GLframer-motion
@vaidyayash8 YavLabs

Infrastructure

I run my own production cloud

Everything I build runs on infrastructure I operate myself — a two-node Proxmox cluster hosting production-grade services with zero-trust networking.

Two-node Proxmox cluster running production-grade self-hosted services.

Traefik reverse proxy, Authentik SSO, Pi-hole DNS, and n8n automation.

Tailscale mesh VPN + Cloudflare Tunnels; a remote Docker socket over Tailscale for decoupled Traefik deployment.

n8n workflow for natural-language Google Drive search via a Telegram bot.

runningProxmoxTraefikAuthentikTailscaleCloudflaren8nPi-holeDocker

Stack

Tools I build with

Frontend

ReactNext.js 14TypeScriptViteTanStack QueryTailwindshadcn/uiThree.js

Backend

FastAPIPythonNode.jsExpressSQLAlchemyAlembicPrismaGo

Event-Driven & Async

Kafka / RedpandaRabbitMQCeleryBullMQarqMQTTTransactional Outbox

AI / ML

RAGOpenAIAnthropicLangChainOllamapgvectorOpenCVYOLO

Data & Storage

PostgreSQLTimescaleDBpgvectorMySQLMongoDBRedisMinIO / S3

Cloud & DevOps

AWSLambdaStep FunctionsDockerKubernetesProxmoxTraefikCI/CD

Security & Auth

RBAC + ABACOAuth / OIDCSAML 2.0WebAuthnMFA / TOTPAES-256SSH CA

Self-Hosting

AuthentikTailscaleCloudflaren8nPi-holeGrafana / LokiNginx

“He hears a customer concern on a Monday call and is closing the loop in production by Wednesday — without a handoff, without a translation layer, and without losing the original intent.”

“He owned the architecture, the procurement, the build-out, and the cutover, and he did it without disrupting a single client engagement.”

If we had to rebuild our manufacturing-technology practice from scratch tomorrow, Yash is the first person I would call.

Director of Products
ITHENA

Contact

Let's build something great

Hiring, collaborating, or just want to talk shop? Send a message — it lands straight in my inbox.

Email
[email protected]
LinkedIn
/in/yash-vaidya
GitHub
@vaidyayash8
Location
Richmond, VA