Senior GTM AI Analyst

Overview

You're building agentic AI workflows that turn Rubrik's data into actionable intelligence for their sales organization. This means you're writing queries, designing automation, training models, and working with sales leadership to figure out which signals actually predict deal velocity. You're working at a ~5,000-person public company selling enterprise data security and backup solutions, so you have mature data systems but also legacy infrastructure to work around.

Role Snapshot

Company Context

Stage: Public (IPO'd in 2024)

Size: ~4,900 employees

Growth: Mature growth stage - focused on efficiency and AI transformation

Market Position: Leader in data security and backup/recovery; competes with Veeam, Commvault, Cohesity

What This Role Actually Is

This isn't traditional Rev Ops reporting. You're building the next generation of sales tools - think AI agents that surface which accounts to prioritize, predict deal risk, recommend next actions, or automate data enrichment. Rubrik has the data infrastructure (Salesforce, Gong, product usage telemetry, billing systems) but needs someone to wire it together into intelligent workflows.

You're working at the intersection of:

• Data engineering: Writing SQL, building pipelines, maintaining data quality
• Machine learning: Training predictive models, evaluating AI agent performance
• GTM strategy: Understanding what sellers actually need vs what sounds cool
• Product thinking: Building tools people will actually use daily

The "agentic" piece means you're not just creating dashboards - you're building systems that recommend or take actions automatically.

What You'll Actually Do

Time Breakdown

Building/Coding (40%) | Stakeholder Collab (25%) | Analysis/Research (20%) | Maintenance (15%)

Key Activities

• Pipeline Development: Writing Python scripts and SQL queries that pull data from Salesforce, product analytics, Gong call transcripts, and other sources. You're cleaning, joining, and transforming data to feed AI models and automation workflows.

• Model Training & Testing: Building and evaluating ML models for things like lead scoring, deal risk prediction, or account prioritization. You're testing accuracy, tuning parameters, and figuring out which signals actually matter versus which are just noise.

• Workflow Design: Working with sales leaders and reps to understand their actual pain points (not what they say they want, but what slows them down). Then designing agentic workflows - automated alerts, recommended next steps, or data enrichment that happens in the background.

• Tool Integration: Connecting your AI outputs back into the tools sellers actually use - surfacing insights in Salesforce, Slack, or whatever interface makes sense. A model that lives in a notebook is useless; it needs to be where people work.

• Stakeholder Translation: Explaining technical concepts to sales VPs who don't care about your model architecture, and translating vague sales requests ("we need better insights") into concrete data problems you can solve.

• Measurement & Iteration: Tracking whether your tools actually get used and whether they impact outcomes. Did reps who got your account prioritization score close more deals? Did your deal risk alerts lead to faster intervention?

The Honest Reality

What's Hard

• Data Quality Hell: Enterprise data is messy. Salesforce fields aren't standardized, product usage data has gaps, and you'll spend significant time just getting data into a usable state before you can do anything interesting with it.

• Adoption Friction: Building a cool AI tool is one thing; getting 500 sellers to actually use it is another. Most of your workflows will get ignored unless you nail the "last mile" of delivery and change management. Reps already have 10 tools they don't use.

• Ambiguous Requirements: Sales leaders will say "we need AI" without knowing what problem they're solving. You have to be part consultant - digging into what actually drives performance at Rubrik versus chasing whatever sounds impressive in the moment.

• Legacy Tech Debt: Public companies have layers of old systems, custom integrations that barely work, and "that one thing we can't touch because no one knows how it works." You'll hit roadblocks where the data you need exists but is locked in some ancient database.

• Proving ROI: It's hard to isolate the impact of a tool when deals have 10 influencing factors. You'll be constantly defending whether your work actually matters or is just "nice to have."

What Success Looks Like

• Sellers using your tools daily without being reminded - it's embedded in their workflow
• Measurable impact on deal velocity, win rates, or pipeline quality (even if directional)
• Sales leadership asking you "can we do this for X?" - they trust you solve problems
• Your models staying accurate over time without constant retraining
• Other teams (marketing, CS) wanting to use your frameworks for their own workflows

Who You're Working With

Internal Stakeholders:

• Sales VPs & Directors: They own outcomes but don't know what's technically possible with AI. You translate their fuzzy goals into data problems.
• Rev Ops / Sales Ops: Traditional reporting teams who maintain Salesforce hygiene and dashboards. You're building the "next gen" layer on top of their foundation.
• Data Engineering: They own core pipelines and warehouses. You'll request access, propose new data sources, or escalate data quality issues.
• Sales Reps & Managers: End users of your tools. You need their feedback on what's actually useful versus theoretical.
• Product/Engineering: If you're using product usage data to score accounts, you'll work with product analytics teams.

What They Care About:

• Sales leaders: "Will this help us hit our number?"
• Reps: "Does this save me time or make my life harder?"
• Rev Ops: "Will this break our existing reporting?"
• Data teams: "Is your request feasible and maintainable?"

Requirements

• Strong SQL & Python: You're writing production-quality code daily. This isn't analyst SQL; you're building pipelines and automation.
• ML/AI Experience: Familiarity with training models, evaluating performance, and understanding when AI is the right tool (vs when a simple rule works fine).
• GTM Domain Knowledge: You need to understand how B2B sales works - deal stages, pipeline math, what "good" activity looks like. Can't build useful tools without knowing the domain.
• Stakeholder Management: Comfortable presenting to executives, gathering requirements from skeptical sales reps, and pushing back when requests don't make sense.
• Scrappiness: Public company resources but startup-level ambiguity. You'll be figuring things out as you go, not following a playbook.
• Data Tooling: Experience with Salesforce, data warehouses (Snowflake/BigQuery), BI tools, and ideally some familiarity with AI/ML platforms (Langchain, vector databases, etc.).
• Bias Toward Shipping: You're measured on impact, not perfect models. Can you get something useful into sellers' hands quickly and iterate, versus spending 6 months building the "right" solution?

The Reality of "AI GTM"

This is a new function at most companies, including Rubrik. You're not walking into a well-defined role with clear success metrics. You'll be figuring out what "agentic workflows for sales" even means in practice, what's hype versus helpful, and how to build credibility for a new team.

The upside: You're shaping something from scratch at a large company with resources and data. The downside: Unclear scope, potential political friction with existing Rev Ops teams, and the risk that after 12 months leadership decides "AI GTM" was a buzzword experiment.

You'll spend less time on traditional analytics ("how did Q3 perform?") and more on building predictive/prescriptive tools. But you still need to prove value in a language sales leaders understand - pipeline, bookings, efficiency.