Founding Engineer
SelfAgenticMulti-agent AI orchestration
Multi-agent AI orchestration platform in closed beta: AI agents that delegate work to each other, share memory, stream output token by token and act inside tools like Slack, Notion and email.
01
The problem
A single AI assistant answers one prompt at a time. Real work spans several steps and several tools, so it needs agents that can hand work to each other, remember what happened before and act inside the tools a team already uses.
02
What I built
I am a founding engineer on SelfAgentic, a multi-agent AI orchestration platform now in closed beta at selfagentic.in. Agents reason on their own, delegate work to each other, share contextual memory and stream output in real time, token by token, with integrations into the tools teams already use.
- 01Agents reason on their own, delegate to each other and share contextual memory
- 02Real-time streaming with token-level granularity
- 03Slack, Discord, Notion and email integrations with secure credential management
- 04Scheduled agent runs for recurring workflows
03
Stack
- Next.js
- TypeScript
- Node.js
- LLM Integration
- Redis
- Agent Frameworks
- AI Systems
04
Questions
What is SelfAgentic?
SelfAgentic is a multi-agent AI orchestration platform in closed beta at selfagentic.in. Teams of AI agents call on each other, remember past runs and act inside real tools like Slack, Notion and email.
What was Harsh Rastogi's role at SelfAgentic?
He is a Founding Engineer on SelfAgentic, working on the platform where agents delegate to each other, share memory, stream output in real time and connect to the tools teams already use.
Which tools can SelfAgentic agents work with?
Slack, Discord, Notion and email, with credentials stored securely. Agents can also run on a schedule for recurring work.