Skip to content
Pedro Lacerda
AI GTM Engineer · Revenue AutomationOpen to roles

I build AI-powered systems that turn outbound activity into measurable pipeline.

I combine full-cycle sales experience with AI, automation, CRM and data to build GTM systems that reduce manual work and create measurable pipeline.

Explore the AI Sales AgentView my experienceDownload CV
Pedro Lacerda, AI GTM Engineer
live workflowlive
Atlas Analytics
Sofia Martins · VP of Sales
High
ICP score87 / 100
Signal
Hiring 12 Account Executives
Message readyCRM synced
84
partner contracts closed
25.1%
reply rate · 1,758 leads
148%
peak quota attainment
The system

A revenue engine, operated end to end

From a raw signal to booked revenue — the pipeline every GTM system I build follows, with each stage instrumented and measurable.

Revenue Engine Mapsignal → revenue
01/ 08

Signal

A lead source or buying signal enters the system — inbound, list, or trigger.

The product

The AI Sales Agent

An autonomous WhatsApp SDR agent I built and run in production — a deterministic guardrail mesh at its core, with a human in the loop on the contract. It captures, qualifies, converses, books and syncs, end to end.

01/ 07

Capture

Inbound and outbound leads enter the system.

View the AI Sales Agent
The problem

The GTM problems I solve

01

Salespeople spend more time operating tools than selling

Account research, lead qualification, message writing, follow-ups, booking and CRM updates happen across disconnected systems.

I turn repetitive steps into one coordinated workflow.

02

Outbound volume grows while quality collapses

Generic automation creates activity, not pipeline.

I combine account context, qualification logic and personalized messaging to increase volume without removing commercial judgment.

03

CRM data exists, but nobody knows what to do next

Records are incomplete, follow-ups disappear, and teams struggle to prioritize the right accounts.

I build workflows that turn CRM data into concrete actions.

04

AI experiments never become part of the sales process

A prompt is not a GTM system.

I build around real sales behavior, measurable outcomes, human oversight and continuous feedback.

Proof

System scale and commercial proof

System scale
1,758
Leads processed
Contacted by the AI SDR agent in a single 31-day production run.
11,925
Messages processed
2,014 autonomous responses at an 8s median (11s p90).
Commercial proof
84
Partner contracts
Closed-won by the agent during the 31-day run.
25.1%
Reply rate
Across the 1,758 leads contacted (441 replies).
148%
Peak monthly quota
Best month at ClassPass; 120%+ sustained.
#1
AE company-wide by MRR
Two consecutive months at Amenitiz, at 108% average quota.

Every figure matches the downloadable CV. Agent metrics are from a single 31-day production run; results vary by period, market and context.

Selected work

Systems I've built and run

All work
Featured case

AI Sales Agent

A production, serverless WhatsApp SDR agent — a deterministic guardrail mesh at its core, not an LLM wrapper.

Problem
Manual research, qualification, personalized outreach and follow-up don't scale on WhatsApp — and generic automation produces activity, not pipeline.
Built
An inbound + outbound agent on Cloudflare Workers: multi-model routing, a guardrail mesh, Whisper voice, Calendly booking, RAG over a playbook and a React operator console — human-in-the-loop on the contract.
Result
84 partner contracts and 11,925 messages in a 31-day production run; 25.1% reply rate.
Cloudflare WorkersDeepSeekAnthropicTwilio WhatsAppCalendlySupabaseTanStack Start
Explore the case study
Additional system

Clay Enrichment & Lead Scoring

A waterfall enrichment and scoring pipeline with AI personalization over live prospects.

Problem
Single-provider enrichment leaves gaps and generic messaging; the pipeline needed a high match rate at controlled per-record cost.
Built
A waterfall over multiple data providers with fallback logic to maximize match rate while controlling cost, feeding an AI-generated personalization column from firmographic and signal data.
Result
200+ live prospects enriched and scored with per-prospect messaging.
ClayWaterfall enrichmentAI personalizationFirmographic + signal data
See the stack
Experience

Nine years carrying quota

Full experience

ClassPass

Field Account Executive I · Jan 2026 – Present

Full-cycle partner acquisition in Brazil at 120%+ quota, peak 148%.

Amenitiz

SDR → Account Executive · Jul 2023 – Sep 2025

#1 AE company-wide in MRR for two consecutive months; promoted from SDR in 7 months.

Vento Fresco

SDR → Account Manager → Head of Sales · May 2017 – Jul 2023

From 80+ cold calls a day to leading a 30-rep organization to 30% year-over-year growth.

Capabilities

What I build with

Organized by capability, not a logo wall.

Full stack

Automation

  • Workflow orchestration
  • Branching
  • Retries
  • Error handling
  • Monitoring

AI

  • Account research
  • Classification
  • Personalization
  • Structured outputs
  • Human approval

CRM Systems

  • Lead routing
  • Opportunity management
  • Salesforce workflows
  • Data synchronization
  • Pipeline visibility

Data

  • Validation
  • Deduplication
  • Scoring
  • Analytics
  • Attribution

Sales Execution

  • Outbound
  • Discovery
  • Negotiation
  • Closing
  • Partner onboarding
Tools: Cloudflare Workers · Supabase / PostgreSQL · Salesforce · HubSpot · Clay · Twilio WhatsApp · Calendly · DeepSeek · Anthropic · TanStack Start · Whisper (Workers AI)
FAQ

What recruiters usually ask

Yes. I defined the problem, designed the architecture, wrote the deterministic guardrail mesh, and run it in live production on Cloudflare Workers — with a human in the loop on the final contract. In its first 31-day run it closed 84 partner contracts across 1,758 leads.

I ship production systems: multi-model routing, RAG, guardrail design, Cloudflare Workers, Twilio, the Calendly API and Supabase — I read, adapt and ship with AI assistance. My focus is applied technical execution, not framework trivia. I'm happy to walk through the architecture and code in detail.

AI GTM Engineer, GTM Engineer, Revenue Automation, Growth Engineering, and technical GTM roles combining system building with commercial execution.

I am based in Rio de Janeiro, available for global remote work, and hold Brazilian and Portuguese citizenship.

Portuguese native, English fluent, and Spanish fluent.

Let’s talk

Let’s build a better revenue engine.

I’m interested in roles where AI, automation, data and commercial execution come together to create measurable pipeline.

Book a 20-minute conversationEmail PedroDownload Resume