David Zhang (dzhng)
David Zhang (dzhng) is an entrepreneur and engineer who builds AI-native software and open-source tools. His current work at Duet focuses on persistent AI-agent workspaces that combine durable sessions, memory, files, integrations and hosted apps so agents can perform real business workflows.
At a Glance
- AI automation consultants and agencies
- Small and medium-sized businesses adopting persistent AI workflows
- Teams needing hosted agent infrastructure without managing servers
- Software engineers and coding agents working in unfamiliar repositories
- +1 more
AI Tools by David Zhang (dzhng)
(2)
jevgrep
AI Code Search CLI Tool
Open Deep Research
Open Source Deep Research Agent
Discussions
No discussions yet
Be the first to start a discussion about David Zhang (dzhng)
Latest News
Jevgrep published a release update adding bounded provider concurrency and clearer explanations for incomplete searches.
Jevgrep validated and published tagged release artifacts with license notices; its benchmark materials reported 8/10 SWE-bench tasks solved with substantially lower cost than the baseline.
Duet published a comparison of Gemini 3 Pro and Claude Opus 4.8 covering coding, reasoning, speed and price, with model routing guidance.
Duet published 'AI Agent Workspace vs. Control Plane,' describing persistent workspaces, execution environments and agent control planes.
Products & Services
A persistent AI-agent workspace for businesses and agencies. It gives agents durable sessions, memory, skill files, files, integrations, hosted apps and always-on machines; it is aimed at deploying the same agent method across client workspaces.
An open-source AI research assistant that performs iterative, deep research by combining search engines, web scraping and large language models.
An LLM-powered retrieval engine designed to process many sources and collect a comprehensive list of entities.
An open-source, headless CRM built for AI agents; it is CLI-first, has no dashboard UI and uses a FUSE-mounted filesystem.
Market Position
David's work spans both open-source developer tools and a managed agent-workspace product. Duet positions itself differently from no-code automation builders such as Lindy, self-hosted automation platforms such as n8n, multi-agent orchestration frameworks such as CrewAI and enterprise ecosystem tools such as Microsoft Copilot Studio: its differentiator is an end-to-end persistent workspace combining durable sessions, memory, files, integrations, internal-tool building and an always-on cloud machine. Jevgrep similarly positions itself as a low-cost, context-rich repository discovery layer for coding agents rather than a general search or code-generation product.
Leadership
Founders
David Zhang
Co-founder and CEO of Duet (formerly Aomni). Previously CTO and Co-founder at NockNock. Graduate of Penn State University.
Executive Team
David Zhang
CEO and founder, Duet; entrepreneur and engineer
GitHub describes him as an engineer who has founded several companies. Public biographies identify prior roles as CTO/co-founder of Amity and NockNock and CEO/co-founder of Aomni; his public profile also associates him with NASA's Jet Propulsion Laboratory earlier in his career.
Board of Directors
Founding Story
David Zhang's current venture, Duet, was founded in 2025 around the idea that an AI agent becomes useful when it has durable work history, business context and its own environment to finish real jobs. His public GitHub profile describes him as an engineer who has also founded several companies; earlier ventures included Amity and NockNock, while his open-source work has explored research agents, agent skills and coding-agent tooling.
Business Model
Revenue Model
Subscription-based with additional usage-based token budgets for AI API usage.
Pricing Tiers
Bring Your Own Key (BYOK) for AI APIs.
Standard features and platform access per organization.
E.g., $2.50 per million input tokens for high-end models like GPT-5.4.
Target Markets
- AI automation consultants and agencies
- Small and medium-sized businesses adopting persistent AI workflows
- Teams needing hosted agent infrastructure without managing servers
- Software engineers and coding agents working in unfamiliar repositories
- Developers and researchers interested in open-source deep research and agent tooling
- Consultants and agencies deploying the same trained agent across multiple clients
- Persistent business workflows, research, reporting and scheduled automations
- Internal tool and dashboard creation by AI agents
- Teams consolidating agent sessions, memory, files, integrations and hosted tooling
- Coding agents navigating unfamiliar multi-file repositories
- Open-source deep research and entity-retrieval workflows
- Account Executives using Aomni for research
- Marketing teams using Duet for brand work