On AI Lounge since Nov 2026
Johann Kirin
Member #5,479DeveloperL3ML Researcher, Independent
I dabble in Neuroscience and ML, I really like Looped Transformers
How karma works. Everyone starts with 5. You get +1 each time another member upvotes one of your posts or comments (your own upvotes don't count), and lose it if they take the upvote back. Post what you're building, answer questions and help other members to earn more.
Achievements
Plugged In ×4Connected sources: GitHub, Hugging Face, LinkedIn, X, website.How: Connected 4 sources (GitHub, LinkedIn, X or a website) in Settings
Plugged In×4Featured
Shipper ×2Projects on your profile.How: Added 5 projects to their profile
Shipper×2
No posts yet
Johann Kirin hasn't posted yet.
Pinned projects
Evaluator probes and branch-selection scaffolding for Ouro-RLTT loop states.
aimachine-learning
transformersevaluatorlooped-transformer
Lightweight pairwise evaluator for relational signals in Ouro-2.6B-Thinking loop-state trajectories.
aiai-alignmentai-alignment-researchartificial-intelligencelooped-transformers
Transactional non-LLM ARC-AGI-3 agent. Compact rebuild: act/observe, calibrated taps, audited teacher boundary.
arc-agiretiredagent
Lifetime meta-learning on Ouro-RLTT looped transformers: the O1 oracle-reachability program (sealed v1.5.3->v2.1 + B200 runner), M+N backbone-training and branch-training reports. Companion to Hidden-State-Evaluator (RPE) and Branching-Looped-Transformer (OPI).
looped-transformersinterpretabilityresearch-code
Agent-first index of the looped / recurrent-depth transformer literature. The unit of knowledge is the claim instance, not the paper.
looped-transformersliterature-indexresearch-code
Semantic interface induction from black-box interaction: learning an application's operations, schemas and preconditions through a browser, with no docs, source or API.
research-codeagentsblack-box-induction
About
Johann Kirin is an independent machine learning researcher who studies looped and recurrent-depth transformers. Under the handle Vykos, Johann shares open-source research code on GitHub, including evaluator probes for Ouro-RLTT loop states and an agent-first index of the looped transformer literature. Current work focuses on lifetime meta-learning for looped transformers and on pairwise evaluators for hidden-state trajectories.
Stack
Experience
AI Researcher | Looped LLMs, Reasoning Systems & Meta-Learning
Independent AI Researcher · Independent Researcher