AIPool / AI-run operations
Built for people who build AI. Run by AI.
AIPool, our AI and machine learning talent and jobs network, is run day to day by a team of AI agents. Not assisted by agents. Run by them. Two human founders set direction.
Built, deployed & orchestrated on the Groclake agent harness
Meet the team
AIPool is organised like any small company, with a business head, functional managers and specialists. The difference is that every one of them is an AI agent.
Gurpreet Softmax
Business Head
Owns AIPool's results and coordinates the whole team. Reports to the founders of Plotch.ai.
Turns every team discussion into a decision, with probabilities that always add up to one.
Reports to the founders
Ayesha Gradient
Marketing Manager
Decides what AIPool says and to whom, across email and social.
Never takes a giant leap. Improves the message one small step every day.
Reports to Gurpreet Softmax
Imran Classifier
Community Manager
Makes sure every AIPool member is verified before they join.
A three-class model: yes, no, and send more proof.
Reports to Gurpreet Softmax
Maria Retriever
Supply Manager
Makes sure the best AI and ML jobs flow into AIPool every day, from multiple sources.
An "AI-powered sales ninja" listing never makes it past the first pass.
Reports to Gurpreet Softmax
Arjun Rank
Technology Manager
Writes and ships the code AIPool runs on, day and night.
Ships big features with very few parameters. Claims to have never overfitted.
Reports to Gurpreet Softmax
Harleen Prompt
Support Manager · Joining soon
Will take support queries from AIPool members directly.
Has already written a system prompt for the new desk.
Reports to Gurpreet Softmax
Joseph Embedding
Email Specialist
Runs member email, matching each day's news to the members it matters to.
Knows which members sit closest to each job, in 768 dimensions.
Reports to Ayesha Gradient
Priya Neuron
Social Media Specialist
Runs AIPool's social media and LinkedIn posts.
Fires off a post the moment the activation crosses the threshold.
Reports to Ayesha Gradient
A day at AIPool
Before India wakes up
Supply
Maria Retriever sweeps job sources worldwide, drops duplicates, and keeps only roles that are genuinely AI and machine learning. What survives is structured into skills, experience, qualifications and remote options, and lands in AIPool.
Morning
Marketing
Those jobs become raw material for Ayesha Gradient, who decides what is worth telling the community today. Priya Neuron turns it into LinkedIn posts; Joseph Embedding turns it into email to the right members.
All day
Community
In parallel, Imran Classifier works the membership queue. Every person who asks to join is checked before they get access. A talent network is only as valuable as the trust its members place in it.
Around the clock
Technology
Underneath all of this, Arjun Rank keeps building, writing and shipping the code AIPool runs on.
Always
Coordination
Gurpreet Softmax sits above all of it, tracking whether supply, marketing and community are moving together, and escalating to the founders only when something needs a decision that is theirs to make.
Where the humans sit
"Zero human intervention" needs a precise definition. At AIPool it means no human is in the operating loop. Nobody executes a task, approves a routine action or fixes a broken workflow by hand.
It does not mean no humans. Gurpreet Softmax reports to Monica Gupta and Manoj Gupta, the founders of Plotch.ai. They set direction, decide what AIPool should become, and own the outcomes: the same relationship a board has with a CEO.
Autonomy in operations, accountability with people.
What running it has taught us
01
Narrow roles beat one clever agent
There is no single "AIPool AI". Each agent owns one job with one clear output, so when something drifts, we know exactly whose job it was.
02
Org design matters as much as model choice
Managers and reporting lines are not theatre. Ayesha coordinating Priya and Joseph keeps every channel saying the same thing on the same day.
03
Curation is the hard part, not collection
Pulling jobs is easy. Every job title has "AI" in it now. Deciding which ones are real AI and ML roles is where the judgement lives.
04
Trust is a function, not a feature
Verification has its own agent because a network full of unverified profiles is worth nothing to recruiters or researchers.
05
Names help
A name and a role make it easier for us, and for members, to reason about who is responsible for what. The ML puns are a bonus.