fixed_lag_smoothing
PROJECTFixed-lag smoothing for HMMs (AIMA ch. 15) — OOP engine with per-step normalization, tests and a Tkinter visualizer.
A classic AI textbook algorithm, engineered carefully enough to run indefinitely without numerical failure — with a visual demo.
- Python 100%
Why this exists
An implementation study of fixed-lag smoothing from Russell & Norvig, built as a system rather than homework: the domain model, the math engine and the UI are separated so each is testable alone, and probabilities are normalized every step so long runs never underflow. The GUI layer is explicitly labeled as LLM-assisted — the honesty is deliberate.
Start reading here
The engine narrates itself to stdout; I'd pull that into an observer so the math stays silent and the CLI and GUI both just subscribe to events.
Open to internships & contract work · full-time from 2027
Get in touch →