The Engineer Who Orchestrates Everything
Armita Peymandoust's Twitter bio has one line that does all the work: "Mom, Optimist, Immigrant. Stanford Engineering. Making ML and AI accessible to all." No titles. No buzzwords. Just the facts that actually shaped her. That discipline - cut the noise, show the signal - is what built Agentforce.
Right now she's SVP of Engineering, AI Cloud at Salesforce, which means she's responsible for the infrastructure that thousands of enterprises are betting their transformation stories on. But the path here is the interesting part. She started at the University of Tehran studying electrical engineering, crossed continents to Northeastern for her master's, then landed at Stanford for a PhD that nobody in enterprise software would have predicted would matter. Her dissertation work involved optimizing how MP3 decoders use power - building a tool called OptAlg that automated what previously required a human specialist. That instinct, automating away the specialist bottleneck, runs through everything she's done since.
We are shifting from monolithic AI to a system where a primary 'orchestrator' agent directs smaller, expert agents. This model allows for greater specialization, efficiency, and scalability, much like a well-managed human team.
- Armita Peymandoust, 2026At Intel she designed silicon for the IA-64 processor line - work that requires a different kind of precision than software, one where you cannot ship a patch after launch. Then came Yahoo, where she joined the Smart Ads team and spent her days predicting click-through rates before "machine learning" was a resume bullet point anyone could fake. She was building the real thing: optimization algorithms at scale, feeding real money decisions in near real time.
Salesforce came next, and what happened there is a study in how to compound a career. She joined in product management - a smart pivot for someone who understood systems deeply but wanted to shape what got built, not just how. Over a decade she moved from Senior Director to VP to SVP, tracking Salesforce's own transformation from CRM vendor to AI platform. Each role left a product behind: Einstein Engagement Scoring, Send Time Optimization, Content Tagging, Copy Insights, Einstein Copilot, Agentforce. A portfolio most engineers would kill for; she built it one launch at a time.
The Einstein for Marketing work is worth lingering on. When she announced those capabilities - tools that let marketers tap machine learning without becoming data scientists - she said the quiet part out loud: "I'm excited that we're bringing more machine learning and AI into the hands of marketers, without the need to turn them into data scientists." That framing, AI as a craft tool not a credential, is her consistent argument. She's made it in podcasts, blog posts, conference panels, and now in the architecture of Agentforce itself.
There are a lot of opportunities in AI beyond engineering. 67% of global business leaders are considering using generative AI, but roughly the same number of IT leaders say their employees don't have the skills to use it.
- Armita PeymandoustAgentforce is the current bet. Salesforce's autonomous AI agent platform - where specialized agents handle discrete tasks while an orchestrator manages the workflow - reflects something Armita has been building toward for years. She named it clearly in her 2026 predictions: the "orchestrated workforce." Not AI replacing people. AI as a team that humans supervise, with observability tools to set guardrails and keep ethics in the loop. The framing matters because it's neither utopian nor dystopian - it's operational. Which is exactly how an engineer who also runs teams thinks about it.
One more detail worth keeping: in October 2018, she hosted a Persian Women in Tech event at Salesforce's San Francisco headquarters. SVP hosting a community meetup for underrepresented engineers is not a checkbox move. It's what someone does when they remember being the person who needed that room to exist.