Millnew AI
Explaining time-series models using local LLMs and Captum
Details
- External ID
- 49464732
- Source
- HN
- Company
- —
- Product
- Millnew AI
- Website domain
- streamlit.app
- Launched
- Aug. 27, 2026
- Cohort
- —
- Upvotes
- 5
- Upvotes percentile
- 0.12634408602150538
- Tags
- —
- Fetched at
- Sept. 10, 2026, 5:32 a.m.
- Updated at
- Sept. 10, 2026, 5:32 a.m.
Enrichment
- Theme
- ML inference and model optimization
- Vertical
- Horizontal
- Function
- Observability & eval
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- time-series model explainability with local llms
- Manually corrected
- False
Could you build this?
Partial The UI and local LLM orchestration are straightforward, but correctly computing feature attribution using PyTorch Captum on non-standard time-series deep learning architectures requires ML domain expertise.
What it would actually take: A production version requires a Python backend integrating PyTorch, PyTorch Captum (for Integrated Gradients, DeepLIFT, or SHAP algorithms), and local LLM runtimes like Ollama or vLLM. The challenging aspect is translating gradient-based feature attribution across multi-channel temporal dimensions into structured prompt representations that an LLM can accurately interpret without hallucinating spurious correlations. This demands ML expertise in time-series modeling and model explainability techniques.
Discussion
1 comment analyzed.
Competitors
Other products that read as similar to this one — 830 launches clear the similarity bar, closest 8 shown.
Attention rank: #693 of 831 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 302 days after the earliest competitor.
- Open-source AMDGCN kernels for optimizing LLM inference · hn · 2026-08-25 · 5 upvotes · similarity 0.58
- KiT · github · 2026-09-28 · 10 upvotes · similarity 0.55
- Goku · hn · 2026-07-15 · 9 upvotes · similarity 0.53
- market-echo · github · 2026-09-13 · 29 upvotes · similarity 0.53
- Timber · hn · 2026-03-02 · 207 upvotes · similarity 0.51
- TypeLLM · github · 2026-09-17 · 35 upvotes · similarity 0.51
- gpu-time · github · 2026-09-11 · 170 upvotes · similarity 0.51
- A model index from the AI Gateways · hn · 2026-09-18 · 7 upvotes · similarity 0.51
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a observability & eval tool for Media & entertainment yet.