But that's just semantics.
Let’s start with the obligatory passion statement, because that seems necessary. Machine learning and complex systems provide a dynamic space for transforming abstract ideas into tangible models. In pursuit of knowledge, I'm driven by the endless possibilities of learning (both machine and deep).
The machine is learning, under my supervision (of course). Ironically, I remain confused. Recursion: the cause and solution to all my problems. The Sisyphean struggle means rolling the boulder in search of answers at the base, getting lost in thought, then pushing algorithms up the mountain with each iteration. As a masochist, I spend my time optimizing models that refuse to cooperate—because apparently, I find trial and error entertaining. This is a snapshot of ongoing work and the evolution of my projects over time. Most code gets rewritten, some ideas never make it past their first push—an ongoing hypothesis test proving that not every idea deserves a second look.
High confidence, very predictable
- Language models & NLP: Retrieval, embeddings, multilingual systems, making models understand nuance (or at least fake it)
- Optimization & adaptive systems: Search, reinforcement, algorithms that learn and occasionally refuse to behave
- Uncertainty & complexity: Finding patterns in noise, causal inference, making sense of what isn’t obvious
Not a fixed stack, just the things I use (or break) the most
- Machine Learning & AI: Deep learning, transformers, probabilistic models, graph-based methods
- NLP & Language Models: Fine-tuning, semantic search, multilingual modeling
- Optimization & Experimentation: Bayesian tuning, reinforcement learning, active learning
- Data & Systems: Cloud pipelines, distributed processing, workflow automation
If nothing else, the name is accurate. Just me, the boulder, and the incline.
Perpetually uphill,
Stochastic Sisyphus