Avinash Rayudu · GTM Engineering

5 AI agents running go-to-market work end to end. In production, not demos.

Voice, enrichment, outbound, invoicing, alerting, and the systems underneath them: built, shipped, and running at a multi-channel CPG brand in New York.

Brooklyn, NY · open to GTM engineering, revenue operations, and growth roles

Agents in production

200+ hrsof manual dialing and data entry replaced

Outbound voice agent

Conducts outbound voice conversations with DTC brands and retail partners, qualifies them, and captures the account data the team previously collected by hand. Structured output lands in the CRM, not a transcript dump.

Retell AIElevenLabsHubSpot
In production · running against the live retail and DTC account list
$20K+collected faster into the account

Invoicing agent

Runs the receivables cycle end to end: issue, mismatch flags against PO dates, and follow-up, all before anything reaches finance processing. Cut payment cycles by 15 days and eliminated recurring billing discrepancies, along with the manual invoicing workload.

n8nShopifyFinaloop
In production · runs every billing cycle
1,000+personalized emails sent, leads generated

Enrichment & outbound pipeline

Sources and enriches contacts, verifies emails, personalizes every message individually, and sends scheduled campaigns unattended. Every send is sourced and verified through Clay before it goes out.

Clayn8nHubSpotGmail API
In production · campaigns queue and send without me
60% → 95%on-time delivery

Operations alerting layer

Monitors the order-to-ship workflow across four fulfillment channels and flags exceptions before they reach the customer. Fulfillment errors down 40%.

Zapiern8n
In production · watching every order, daily
100sof roles sourced and scored

Job search agent

Sources roles daily, scores each against a weighted rubric (skills, comp, location, visa safety), and drafts the outreach. Built because I wanted it to exist; now sold to other job seekers.

ClaudeLLM APIsn8n
In production · daily, for me and for customers

How I decide what to automate

A few months into Zacao I found 42% of our SKUs were blocked from fulfillment by unresolved stock flags. Nobody was tracking it. Nobody asked me to look. Product that should have been moving was sitting still. I resolved it before it reached leadership reporting, then rebuilt the order-to-ship workflow with documented SOPs and automated alerts so it could not happen again.

The tooling is the easy half. The job is deciding what is worth automating, then building it so it stays fixed without you.

Stack

AI & Agents
ClaudeCodexRetell AIElevenLabsLLM APIsagent orchestrationprompt design
GTM Engineering
ClayHubSpot + APIsn8nZapierGmail APIenrichment pipelineslead routing
Data
SQLPythonSnowflakeMetabaseLookerPower BITableauGoogle Analytics
Platforms
ShopifyAmazon FBAFinaloopFirebaseNotionJira

Background

Zacao Chocolate2026 – present

Growth operations & GTM engineering. Multi-channel CPG across DTC, Amazon, wholesale, and B2B; built the agent and reporting layer the team runs on.

Virxon · Founder2022 – 2024

B2B SaaS building custom web and mobile applications. Zero to 20+ clients and $10K revenue in six months, 30 people, no outside funding.

NYU TandonMS, May 2026

MS Management of Technology. Capstone: led a 4-person team through a 15-week client engagement across 6–8 market segments, zero scope creep, 100% on time.

Contact

Happy to walk through any of these systems live rather than describe them: the n8n canvas, the Retell dashboard, the Clay table.