Dr. Elena Park
Research Scientist · DeepMind
FOCUS AREAS
LLMs, Alignment, Research
CONNECT_
TECHNOLOGY
ARTIFICIAL INTELLIGENCE_
Systems that reason, act and build alongside us. From Agentic AI to Responsible AI, this is where AI builders, researchers and operators go deep — through focused circles, live events and hands-on challenges.
TOPICS
WHAT THIS COVERS_
The 7 areas active in Artificial Intelligence right now — pick one to see what it's actually about.
Agentic AI
Systems that plan, act and use tools with minimal supervision.
Generative AI
Models that create text, image, audio and code from a prompt.
LLMs
Large language models — architecture, training and how they reason.
AI Infrastructure
The compute, data pipelines and serving layer models run on.
Robotics
Machines that sense, decide and act in the physical world.
Computer Vision
Teaching systems to read, recognise and reason about images.
Responsible AI
Fairness, safety and accountability as systems gain autonomy.
DEEP DIVE
WHERE AI
GETS DEEP_
Seven frontiers the AI community is working through right now — what each one is really about, and the questions worth bringing into the room.
AGENTIC AI
Agents are moving from demos to real work
The hard part of agentic AI isn't getting a model to plan. It's keeping it reliable when tools fail, goals shift and nobody is watching. This is where builders compare notes on evaluations, guardrails and what actually survives production.
- PLANNING
- TOOL USE
- MEMORY
- MULTI-AGENT
- EVALS
- GUARDRAILS
GENERATIVE AI
Generation is easy. Taste is the hard part.
Anyone can prompt an image or a paragraph now. The real work is control, provenance and building products people come back to — not outputs that only impress once.
- DIFFUSION
- MULTIMODAL
- CONTROL
- PROVENANCE
- CREATIVE TOOLS
- PRODUCT
LLMS
Understanding the model beats tuning the prompt
Context length, retrieval, fine-tuning and reasoning all trade off against cost and latency. Going deep means knowing why a model behaves the way it does, not just which prompt happened to work last week.
- RETRIEVAL
- FINE-TUNING
- REASONING
- CONTEXT
- INFERENCE COST
- OPEN MODELS
AI INFRASTRUCTURE
Every breakthrough runs on someone's infrastructure
Compute, data pipelines and serving decide what's possible long before a model ships. It's the least glamorous layer of AI, and the one that most often decides who gets to build at all.
- GPUS
- TRAINING
- SERVING
- DATA PIPELINES
- MLOPS
- EFFICIENCY
ROBOTICS
Intelligence gets real when it has to move
Robots act in a world that doesn't reset. Simulation, sensing and control meet messy physics, and the gap between a lab demo and a machine that works every day is where the real learning happens.
- CONTROL
- SIM-TO-REAL
- MANIPULATION
- SENSING
- AUTONOMY
- HARDWARE
COMPUTER VISION
Seeing is still one of AI's hardest problems
From medical scans to factory floors, vision systems have to be right in conditions their training data never showed them. Robustness, labelling and running on the edge matter as much as the accuracy score.
- DETECTION
- SEGMENTATION
- EDGE AI
- ROBUSTNESS
- DATASETS
- VIDEO
RESPONSIBLE AI
Trust is built into systems, not added after
Fairness, safety and accountability are engineering problems as much as policy ones. The people shipping AI and the people governing it need to be in the same room — and early.
- SAFETY
- FAIRNESS
- ALIGNMENT
- POLICY
- TRANSPARENCY
- ACCOUNTABILITY
COMMUNITIES
THE CIRCLES IN THIS DOMAIN_
Communities around shared interests in AI. Join one to swap ideas, ask questions, and find people working on the same problems.
DISCUSSION
OPEN QUESTIONS IN AI_
The questions the AI community is arguing about right now — alignment, open models, data and what actually ships. Pick a thread and bring your view.
EVENTS
WHAT'S HAPPENING IN AI RIGHT NOW_
Events, talks, and gatherings.
NOTHING UPCOMING IN THIS DOMAIN YET_ — check all events.
THE NETWORK
THE PEOPLE BUILDING IN AI_
Researchers, founders, builders and investors in AI. Filter the network and find the people worth meeting.
Marcus Chen
Founder · Vector AI
FOCUS AREAS
Agentic AI, Infrastructure
CONNECT_
Priya Desai
Investor · a16z
FOCUS AREAS
AI x Society, Startups
CONNECT_
Jordan Lee
ML Engineer · Anthropic
FOCUS AREAS
Robotics, Computer Vision
CONNECT_
CHALLENGES
THE AI ARENA CHALLENGES_
Challenges for real-world impact.
ONLINE · GLOBALLIVE
AGENTIC AI × RESPONSIBLE AI
Build an AI Agent for Scientific Discovery
2 WEEKS LEFT
ONLINEJUDGING
RESPONSIBLE AI
Red-team an agent against prompt injection
JUDGING
ONLINELIVE
AI INFRASTRUCTURE
Cut inference cost for a 7B model at spiky traffic
9 DAYS LEFT
HYBRIDOPENS SOON
GENERATIVE AI
An evaluation suite for multimodal generation
OPENS 12 MAYELSEWHERE
OTHER RELATED DOMAINS_
Explore other areas at HATCON.
YOUR NEXT MOVE
READY TO GO DEEP IN AI?_
Join the circles, questions and challenges shaping Artificial Intelligence.
