This entry collects smaller AI experiments from 2018 to current where the main value was testing a technique, exploring a workflow, learning a model family or proving whether an idea was useful enough to continue. The work has included algorithmic trading experiments, sentiment analysis, NLP workflows, reinforcement-learning agents learning game environments, drug-discovery-adjacent ideas and many other applied-AI tests.
2018—Current
Period
Experiments
Output
Mixed AI
Scope
Key details
- Explored algorithmic trading ideas and market-signal experiments.
- Worked on sentiment analysis and NLP workflows for extracting useful signals from text.
- Tested reinforcement-learning agents learning game environments and simulations.
- Built smaller drug-discovery and molecular-AI experiments alongside the larger dissertation pipeline.
- Used many smaller experiments to build intuition before committing to larger systems.
- Focused on practical usefulness, failure modes and iteration speed.