Hi, I’m Aarya Shah. I turn problems into AI systems that actually ship.

AI Product Operations Manager at Tutero. Before I build anything with AI, I ask: should this be an agent, a deterministic automation, or left alone? Open to interesting conversations.

27k+
Namo downloads
6
layers covered
3
companies
01

What I do, broadly.

Business first, technical enough to ship. AI systems only work when they connect to real data, real teams, and real failure modes.
01
Product Operations
The processes and systems behind how Tutero runs, more agents in production, and raising how every team works with AI.
02
AI Systems
MCP connectors, agent infrastructure, RAG, fine tuning, turn detection, prompt engineering, LLM pipelines.
03
Automation
Make, n8n, rules engines, workflow debugging, replay scripts, Slack ops, CRM sync, Stripe automation.
04
Platform
Cloud functions, Kubernetes manifests, internal APIs, Firestore, Snowflake, observability tooling.
02

Toolbox, the honest version.

64 tools · 10 categories
PythonJavaScriptDart / FlutterJavaSQLC / C++Shell
HTML / CSSTailwindAstroFlutter
OpenAIAnthropicGroqHugging FaceTensorFlowPyTorchBERTCLIPFine tuningPrompt engineering
VideoSDKLiveKitWhisperElevenLabsSIPTwilioWebRTCSTT / TTS
LangGraphMCPLangfuseCustom loopsOpenAI FunctionsTool use patterns
FAISSQdrantEmbeddingsSentenceTransformersHybrid search
Maken8nCloud FunctionsWebhooksGraphQLREST
StripeTwilioSendGridSlackMondayNotionDialpadCircle
PostgresFirestoreSnowflakeRedis
AWSDockerKubernetesCloudflareGitHub ActionsGit
Most of this lives behind login walls. The pattern shows up the same in every project.
03

Three companies, same playbook.

Vadodara → Surat → Melbourne
Melbourne, AU · Remote · Full-time
AI Product Operations ManagerNov 2026 → Present
  • Promoted after my first year to own product operations end to end: the processes and systems behind how Tutero runs, more agents in production, and raising how every team works with AI.
AI & Automation EngineerOct 2025 → Nov 2026
  • Made Hermes, our agent server, usable company wide: 1 gateway, 7 Slack workspaces, tracing on all 51 agent profiles.
  • Built MCP connectors for Dialpad, SendGrid and Circle. Parent texts now reach our account manager agent in under 2 seconds, down from 4 to 8.5 hours.
  • Automated churn with a rules engine instead of an agent. n8n selected my talk on that kind of call for In The Loop 2026.
  • Authored the agent migration playbook that answered 14 architecture questions from the founder.
  • Shipped an internal forms platform to replace Typeform in about three weeks.
  • Production automations across Make, n8n, Stripe, Monday, Twilio, SendGrid and Firestore, with webhook and API architectures for real time billing and CRM sync.
Surat, IN · Internship
  • Trained and fine tuned Namo, a semantic turn detection model for voice agents.
  • Shipped to the AI Voice Agents framework: SIP integration, Twilio transfer, Azure OpenAI, STT, TTS and LLM providers, plugin architecture.
  • Production voice agent deployments on AWS. Public demo site and developer documentation.
  • n8n and Docker content workflows on AWS: Reddit, X and LinkedIn syndication via REST.
Shipped publicly
Namo Turn Detector
Semantic turn detection so voice agents know when you’re actually done speaking. 23+ languages, under 30 ms on CPU, 27k+ downloads.
Article ↗HF ↗
Vadodara, IN · Internship
  • Diamond recommendation chatbot: fine tuned LLM, regex, prompt engineering.
  • Fine tuned a generative model for product imagery and integrated OpenAI CLIP for image and natural language search.
✦ Education & Recognition ✦
Education
B.E. Computer Engineering
Gujarat Technological University
October 2021 → June 2025
GPA 8.4 / 10
Selected to speak at
In The Loop 2026
n8n’s first major conference
Deny by default: deciding what your automation is allowed to think.
Selected for
Amazon ML Summer School 2024