Article
1
Kshitij Thakkar PRO
kshitijthakkar
AI & ML interests
AI observability + MoE efficiency engineer. Building tools that make GenAI traceable, measurable, and production-ready.
Recent Activity
updated
a model about 5 hours ago
kshitijthakkar/LFM2.5-1.2B-Instruct-liquidchat-lora-cactus published
a model about 5 hours ago
kshitijthakkar/LFM2.5-1.2B-Instruct-liquidchat-lora-cactus liked
a model about 6 hours ago
Qwen/Qwen3.5-35B-A3B Organizations
Articles 5
Article
1
Systematic Architecture Search for Mobile-Optimized Mixture of Experts Language Models
Systematic search for 1B-2B MoE models. Best: bs=1, ctx=2048 achieves 0.32 loss. Top-8 routing beats top-2.
-
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs4-ctx1024
Updated • 25 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs2-ctx2048
Updated • 26 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs2-ctx1024
Updated • 23 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs1-ctx2048
Updated • 34
Systematic search for 1B-2B MoE models. Best: bs=1, ctx=2048 achieves 0.32 loss. Top-8 routing beats top-2.
-
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs4-ctx1024
Updated • 25 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs2-ctx2048
Updated • 26 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs2-ctx1024
Updated • 23 -
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs1-ctx2048
Updated • 34
32 MoE models from 41 experiments exploring expert count, routing, and learning rates for mobile deployment.
spaces 11
Running
Loggenix MoE 0.4B-A0.2B Demo
🧠
Test and evaluate the Loggenix MoE language model
Running
1
E-Commerce Product Content Generator
🛒
Generate product photos and marketing copy for e‑commerce
Running
1
Multimodal Content Pipeline
🖼
Generate an image and hear its spoken description
Running
1
AI Content Creation Pipeline
🎨
Generate complete social media posts from a text prompt
Build error
2
Loggenix Moe 0.3B A0.1B Demo
🏢
Demo Space for my model loggenix-moe-0.3B-A0.1B
Sleeping
my_Akatsuki_theme
🚀
Preview and customize `my_Akatsuki_theme` in a Gradio interface
models 118
kshitijthakkar/LFM2.5-1.2B-Instruct-liquidchat-lora-cactus
Updated
kshitijthakkar/LFM2.5-1.2B-Instruct-liquidchat-lora
Updated
kshitijthakkar/loggenix-moe-0.4b-cpt-muon-v1
Updated
kshitijthakkar/lfm25-mobile-actions-cactus
Updated
kshitijthakkar/loggenix-moe-0.4b-cactus
Updated
kshitijthakkar/poc-pipeline-e2e-test
Text Generation • 39.4M • Updated
• 50
kshitijthakkar/lfm-finetuned
Text Generation • Updated
kshitijthakkar/LFM2.5-1.2B-Instruct-mobile-actions
Text Generation • Updated
• 88
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs4-ctx1024
Updated
• 25
kshitijthakkar/moe-1083m-781m-16x8-8L-large-moe-1.3b-bs2-ctx2048
Updated
• 26
datasets 382
kshitijthakkar/liquidchat-mobile-actions-2026-02-24
Updated
kshitijthakkar/liquidchat-lora-dataset
Viewer
• Updated
• 32.2k
kshitijthakkar/nemotron-sft-balanced-2b-v1
Viewer
• Updated
• 200k • 5
kshitijthakkar/liquidchat-mobile-actions-2026-02-23
Viewer
• Updated
• 1 • 2
kshitijthakkar/loggenix-0.4B-flagged-responses
Viewer
• Updated
• 5 • 35
kshitijthakkar/large-moe-inference-benchmark
Viewer
• Updated
• 38 • 15
kshitijthakkar/moe-inference-benchmark
Viewer
• Updated
• 256 • 34
kshitijthakkar/nemotron-sft-balanced-stage1-2-ChatML-small
Viewer
• Updated
• 25k • 29
kshitijthakkar/nemotron-pre-training-balanced-250m
Viewer
• Updated
• 113k • 25
kshitijthakkar/nemotron-sft-balanced-stage1-2-ChatML-V2
Viewer
• Updated
• 500k • 28