Archive
01
Cohort 1
The first cohort opened the reading, discussion, and implementation format that later evolved into more structured cohorts.
Open cohortCohort 3, applications active
From frontier papers to AI systems that survive reality.
Tribu IA Papers gathers highly committed technical talent to study recent scientific research, implement it in code, and turn it into applied engineering judgment.
Applications open
Cohort 3
Free, merit-based, and demanding in technical level and commitment. The cohort combines deep reading, technical sessions, applied exercises, and projects with high self-management.
- Rhythm
- 8 to 12 hours per week, several months
- Status
- Applications open
- Focus
- Spring 2026, specialized curriculum
Cohort archive
Cohorts as technical memory
Each cohort has its own focus, activities, achievements, and papers. The trajectory matters: each cohort turns what it learns into cultural infrastructure for the next one.
Completed
02
Cohort 2
The second cohort drew from the Agentic AI MOOC Fall 2025. Starting with this cohort, Leonardo Gonzalez joined as program head.
Open cohortApplications open
03
Cohort 3
The third cohort opens a free, merit-based path for people with strong technical ability, initiative, and real availability.
Open cohortFall 2025, Agentic AI
Cohort 2 raised the bar
The second cohort drew from the Agentic AI MOOC Fall 2025 and pushed the work into competition, demos, and public exposure. Starting with this cohort, Leonardo Gonzalez joined as program head.
- Team participation in UC Berkeley's AgentX AgentBeats.
- One team placed third worldwide in the Gaming category, presented at AI Tinkerers, and received a full scholarship to the Hardcore AI program.
Papers and sessions
Living curriculum
Papers are kept as site data so they can be queried by cohort, topic, or week. Cohort 2 draws from the Agentic AI MOOC Fall 2025; Cohort 3 will reuse part of that material while branching into a specialized curriculum.
LLM Agents Overview
Yann Dubois, OpenAI
Post-Training Verifiable Agents
Jiantao Jiao, NVIDIA
Agent Evaluation and Project Overview
Agentic AI MOOC staff
Predictable Noise in LLM
Sida Wang, Meta
AI Agents to Automate Scientific Discoveries
James Zou, Stanford
Supporting material
General resources
Supporting material for strengthening technical reading judgment outside a specific cohort.
Merit-based selection
Expected profile
- Strong experience in Python, ML/DL, or software engineering.
- Technical reading ability in English and comfort with papers, documentation, and repositories.
- Real availability of 8 to 12 hours per week for several months.
- High self-management, serious collaboration, and ability to produce deliverables.
Applications open
Apply to Cohort 3
Seats are limited, and selection prioritizes merit, initiative, and the ability to sustain deep technical work.
Apply to Cohort 3