Overview
You write code that automates parts of Grafana's go-to-market motion. This includes building AI agents for lead qualification, automating CRM workflows, integrating data sources, and creating tools that help SDRs and AEs do their jobs faster. You report to Tudor Matei (Staff AI Growth Engineer) and work closely with sales ops, marketing ops, and individual reps to understand what's broken and build systems to fix it.
Role Snapshot
| Aspect | Details |
|---|---|
| Role Type | Technical GTM engineer (code-first rev ops) |
| Sales Motion | N/A - you support the sales org, not sell |
| Deal Complexity | N/A - internal engineering role |
| Sales Cycle | N/A |
| Deal Size | N/A |
| Quota (est.) | N/A - measured on shipped systems, adoption, efficiency gains |
Company Context
Stage: Late-stage (1,774 employees, likely Series D+ or approaching IPO based on size)
Size: 1,774 employees
Growth: Actively building out AI/automation capabilities within GTM - hiring at Staff level suggests maturing this function
Market Position: Leader in observability space with strong open-source foundation (Grafana, Loki, Tempo, Prometheus ecosystem). Competing against Datadog, New Relic, Dynatrace in the cloud observability market.
GTM Reality
Pipeline Sources:
- Mix of product-led growth (open-source users converting to Grafana Cloud), inbound demand, and outbound motions
- Large sales org supporting enterprise deals and self-service conversions
- Likely multiple sales segments (SMB, mid-market, enterprise) with different playbooks
Your Role in GTM: You're not selling - you're building the infrastructure that makes selling more efficient. This means understanding sales workflows deeply enough to automate them intelligently.
What You'll Actually Do
Time Breakdown
Coding/Building (50%) | Working with stakeholders (25%) | Testing/Debugging (15%) | Meetings/Planning (10%)
Key Activities
- Build AI agents and automation workflows: Write Python/TypeScript to create systems that score leads, enrich contact data, route accounts, trigger sequences, or surface insights. You're using LLMs, vector databases, and orchestration tools to make sales processes smarter.
- Integrate and wrangle data: Connect CRMs (likely Salesforce), marketing automation (Marketo/HubSpot?), product analytics, data warehouses, and third-party enrichment tools. A lot of your time is spent making sure data flows correctly between systems.
- Build internal tools and dashboards: Create interfaces for sales ops or reps to interact with your automation - think Retool apps, custom Slack bots, or lightweight web apps that surface AI-generated insights.
- Debug and optimize existing systems: When an automation breaks or performs poorly, you troubleshoot. This means digging through logs, understanding why an AI agent made a bad decision, or figuring out why a Zapier-replacement workflow isn't firing.
- Talk to stakeholders about what sucks: Sit with SDR managers, AEs, or ops people to understand manual processes that could be automated. Translate their complaints into technical requirements.
The Honest Reality
What's Hard
- Sales processes are messy: Requirements change constantly. What worked last quarter gets thrown out when leadership changes the ICP or territory model. You'll build things that get deprecated quickly.
- AI outputs are unpredictable: LLMs hallucinate, classification models drift, and "90% accurate" still means you're creating work for someone to review edge cases. You spend time tuning prompts and building guardrails.
- You're building for non-technical users: Sales reps won't read documentation. If your tool isn't immediately intuitive, they won't use it. Adoption is a constant challenge.
- Data quality is terrible: CRM data is incomplete, duplicated, or wrong. Your automations are only as good as the data they process, and you'll spend more time cleaning data than you'd like.
- Measuring impact is fuzzy: Unlike product engineering where you see users or revenue directly, you're measuring things like "time saved" or "lead quality improvement" which are hard to quantify and easy to dismiss.
What Success Looks Like
- Your AI lead scoring model is actually used by SDRs and they trust it enough to prioritize based on its output
- You ship an automation that eliminates 10+ hours/week of manual work for the sales ops team
- Sales leadership references one of your dashboards in QBRs because it's become their source of truth
- Other teams (CS, marketing) ask you to build similar automation for them
Who You're Supporting
Primary Stakeholders:
- Sales Operations / Revenue Operations leadership (your main customers)
- SDR/BDR managers who want better lead routing and activity tracking
- AEs who want cleaner data and less CRM admin
- Marketing ops who need attribution and funnel analytics
What They Care About:
- Speed: Can you ship this automation before the quarter ends?
- Reliability: Will this break when they're demoing it to the VP?
- Simplicity: Can their team actually use this without extensive training?
- Impact: Can you show this saved X hours or improved Y metric?
Requirements
- Strong coding skills in Python and/or TypeScript - you're writing production code daily
- Experience with AI/ML tooling (LLMs, vector databases, prompt engineering, agent frameworks like LangChain/CrewAI)
- Deep understanding of GTM systems and data models (Salesforce, HubSpot, Marketo, CDPs, data warehouses)
- Ability to translate messy business requirements into clean technical solutions
- Track record of shipping automation that people actually use (not just POCs)
- Comfortable working across the stack - APIs, databases, front-end interfaces, orchestration tools
- Experience at a B2B SaaS company where you've seen sales operations up close
- Staff-level expectations: You should be able to define projects, not just execute tickets