About the role
Product Manager — AI Products
BuzzBoard · Remote (WFH) · ProductAbout BuzzBoard
BuzzBoard builds AI products for the B2SMB market — helping agencies, media companies, and sellers understand small businesses and market to them at a level of personalization that wasn't previously economical.
Our platform is built on frontier models end to end. Not AI features bolted onto a legacy product — the products are agent systems.
Our products span multi-agent marketing content generation, real-time AI voice intake, and pre- and post-sales intelligence for SMBs — Zylo, IRIS, Ignite, and Ember. They share a common internal pipeline for orchestration, retrieval, tool use, evaluation, and deployment, and a spec-driven development process that treats agent behavior as a first-class design artifact.
The role
This is a Product Manager role for people who build, not just specify.
Two things are equally true about how we work, and both are non-negotiable expectations of this role:
1. The products you own are agent systems, and they are non-deterministic. Your design surface is prompts, tool definitions, retrieval strategy, schema contracts, model selection, fallback behavior, and failure modes. “The model got it wrong” is a product bug you are expected to diagnose and specify a fix for — not escalate.
2. You use these models to do the product work. You will write specs, generate structured data, produce working HTML prototypes, run evaluation batches, and analyze output quality using the same models the products run on. Our specs ship as machine-readable handoff sets, and prototypes ship with live model stubs. If your definition of a spec is a Confluence page of prose, this role will be uncomfortable.
What you'll actually do
Own one or more product lines — outcomes, roadmap, sequencing, and the quality bar
Write the spec set engineering builds from: product spec, engineering spec, model spec, and handoff readiness — decisions resolved, not deferred
Design the agent behavior — prompt architecture, output schemas, deterministic vs. model-decided logic, guardrails, and what happens when the model is wrong or the tool call fails
Build prototypes yourself — working HTML with live model calls, ahead of engineering commitment, so we argue about a thing instead of a document
Define and run evaluations — build eval sets, score output batches, set ship/no-ship thresholds, and drive prompt and model iteration off measured results rather than vibes
Make the model tradeoffs — model choice, context strategy, latency, token cost per unit of output, fine-tune vs. prompt vs. retrieval
Work directly with engineering, GenAI, and design on daily execution; hold the line on schema and interface contracts between agents
Interface with customers and enterprise partners, including US-based partners with real security, privacy, and compliance review processes
Instrument and read the data — usage, quality, cost, and drift — and act on it
What we're looking for
Required
4+ years in product management, ideally B2B SaaS
Demonstrated hands-on work with LLM / agent products: prompt design, structured output, retrieval, tool/function calling, evaluation. We will ask you to walk through something you shipped in detail.
Fluency using AI tooling in your own workflow — specs, prototypes, analysis, code reading
Ability to read code and API contracts well enough to review a schema, follow a pipeline, and spot a bad interface
Track record of shipping — features live, in customers' hands, with measured outcomes
Strong written communication; you can take a technical decision and make it legible to a CEO and to an engineer in the same document
Comfort operating with ambiguity and short cycles
Preferred
Degree in AI/ML, Engineering, CS, or a related technical field
Direct experience with OpenAI, AWS Bedrock, Anthropic/Claude, LangChain, or equivalent orchestration stacks
Experience with voice AI, agentic workflows, or multi-agent systems
Experience with fine-tuning, dataset curation, or model performance analysis
Experience working with US customers and enterprise partners
SMB or marketing-technology domain knowledge
What this role is not
Not a backlog-grooming or ticket-triage role
Not a role where AI strategy is delegated to a data science team you email
Not a role where a spec can end with “TBD, engineering to decide”
What we offer
Fully remote
Genuine ownership of a product line, with the latitude to shape it
Work at the current frontier of applied AI product development — agent systems, evals, voice, and multi-model orchestration in production
A small, high-context team that moves quickly and argues about the work
Real impact on the small businesses our customers serve
Apply if you're a product manager who builds with these models daily and wants to own products made of them.
Originally posted on Himalayas