cmpnd

Website: cmpnd.ai · Contact: hello@cmpnd.ai · GitHub: cmpnd-ai

Early-stage startup built around DSPy and the “compound AI systems” thesis (chaining models, retrievers, and tools rather than relying on a single model). Positions itself around production-readiness — its tagline: “open-source tools to close the gap between ‘it works on my machine’ and ‘it works in production.‘”

This is the company Heather Miller described to me back in July 2025 as a “Compound AI Systems” company she’d be starting up. Confirmed via research on 2026-06-24.

People

• Jessica Laughlin — CEO / co-founder (named as prospective CEO in the July 2025 intro email; “Drew” in earlier notes may refer to her or a separate person — unconfirmed).
• Isaac Miller — Head of Engineering. DSPy core contributor, ex-Northwestern. Authors much of the cmpnd blog. No relation to Heather Miller. (blog, LinkedIn)
• Heather Miller — CMU assistant professor and VP at Two Sigma; the connective figure who introduced me. Research on DSPy + Language Model System Interface Model.
• Josh Quicksall — Brand Strategy & Communications Lead. Previously comms at CMU’s Software & Societal Systems department. Sponsorship Chair for CAIS 2026.

Product / Output

• dspy-cli — tool to create, develop, test, and deploy DSPy programs as HTTP APIs; auto-generates endpoints, OpenAPI specs, and Docker configs. (pip install dspy-cli / uv). repo
• dspy-tutorial-deep-research — tutorial teaching DSPy core abstractions (Signatures, Modules) by building a multi-step deep research agent. repo
• Blog cadence (production-focused, DSPy-centric):
  • Jan 2026 — Introducing dspy-cli: DSPy Programs as HTTP APIs
  • Jan 2026 — Moderating Discord with DSPy
  • Feb 2026 — DSPy is the easiest way to use RLMs (Recursive Language Models)
  • Mar 2026 — Tutorial: Learn DSPy by Building a Deep Research Agent

Ecosystem position

• cmpnd is a sponsor of ACM CAIS 2026 (inaugural Conference on AI and Agentic Systems, San Jose, May 26–29 2026). Isaac Miller and cmpnd appear on the program committee; Josh Quicksall is Sponsorship Chair.
• Well-embedded in the academic compound-AI community led by Matei Zaharia, Omar Khattab (DSPy creator), Jeff Dean, Ion Stoica, Chris Potts, etc. — the “Shift from Models to Compound AI Systems” crowd.

Funding / Traction (as of 2026-06-24)

• No disclosed funding. No Crunchbase entry, no announced round, nothing in funding trackers. Consistent with pre-seed / early formation / bootstrapped.
• Public signal is open-source tooling + community presence, not revenue or named raise. No public traction metrics (stars/contributors/revenue) gathered yet — TODO if diligence continues.

Viability read (informal)

• Strengths: top-tier technical + academic network; rising category (DSPy adoption growing); credible production-tooling wedge.
• Risks: crowded, largely open-source DSPy tooling space with unproven monetization; small team; no disclosed capital or traction; commercial model not yet externally visible. For an investment there’s essentially no public financial info to underwrite — would need deck, cap table, round terms, revenue directly.

⚠️ Conflict-of-interest flag (important for me)

Heather Miller is listed as VP at Two Sigma — my own employer. Any investment in or move to cmpnd likely implicates Two Sigma’s conflict-of-interest / outside-activities / personal-investment policies and possibly compliance pre-clearance. Separately, a job change carries immigration consequences given my in-progress PERM green-card process. Raise with Two Sigma compliance (and immigration counsel) before further diligence or any commitment.

Open questions / TODO

• Confirm Jessica Laughlin’s exact title and whether “Drew” is a distinct person.
• Pull dspy-cli repo metrics (stars, contributors, release activity) as a traction proxy.
• Check for any newer funding/announcement footprint for cmpnd or Jessica Laughlin.
• Get equity/round details directly if pursuing investment.