I'm Adithya, a 20-year-old dev from India. I have been working with GenAI for the past year, and I've found it really painful to deal with the many different forms of data out there and get the best representation of it for my AI applications.
That's why I built OmniParse\u2014an open-source platform designed to handle any unstructured data and transform it into optimized, structured representations.
Key Features:\n- Completely local processing\u2014no external APIs\n- Supports ~20 file types\n- Converts documents, multimedia, and web pages to high-quality structured markdown\n- Table extraction, image extraction/captioning, audio/video transcription, web page crawling\n- Fits in a T4 GPU\n- Easily deployable with Docker and Skypilot\n- Colab friendly with an interactive UI powered by Gradio
Why OmniParse?\nI wanted a platform that could take any kind of data\u2014documents, images, videos, audio files, web pages, and more\u2014and make it clean and structured, ready for AI applications.
Check it out on GitHub: https://git.new/omniparse"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: I Made an Open Source Platform for Structuring Any Unstructured Data"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/adithya-s-k/omniparse"}},"_tags":["story","author_adithya-s-k","story_40854733","show_hn"],"author":"adithya-s-k","children":[40856481,40859420,40859589,40860846],"created_at":"2024-07-02T08:48:04Z","created_at_i":1719910084,"num_comments":13,"objectID":"40854733","points":88,"story_id":40854733,"story_text":"Hey HN,
I'm Adithya, a 20-year-old dev from India. I have been working with GenAI for the past year, and I've found it really painful to deal with the many different forms of data out there and get the best representation of it for my AI applications.
That's why I built OmniParse\u2014an open-source platform designed to handle any unstructured data and transform it into optimized, structured representations.
Key Features:\n- Completely local processing\u2014no external APIs\n- Supports ~20 file types\n- Converts documents, multimedia, and web pages to high-quality structured markdown\n- Table extraction, image extraction/captioning, audio/video transcription, web page crawling\n- Fits in a T4 GPU\n- Easily deployable with Docker and Skypilot\n- Colab friendly with an interactive UI powered by Gradio
Why OmniParse?\nI wanted a platform that could take any kind of data\u2014documents, images, videos, audio files, web pages, and more\u2014and make it clean and structured, ready for AI applications.
Check it out on GitHub: https://git.new/omniparse","title":"Show HN: I Made an Open Source Platform for Structuring Any Unstructured Data","updated_at":"2024-09-20T17:23:50Z","url":"https://github.com/adithya-s-k/omniparse"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"kordlessagain"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"I have been researching methods and projects built around training and running LLMs locally. I'm interested in what others have been using on this front (including straight up Pytorch/Transformers). Here's what I've gathered so far:
Engines/APIs:
- vllm: Inference and serving engine for LLMs (none quantized models only?) [1]\n - ollama: Go project to run, create and share LLMs [2]\n - llama.cpp: Inference of LLaMA models in C/C++ w/UI (including quantized models) [3]\n - llama-cpp-python: run OpenAI compatible API bindings for llama.cpp [4]\n - llm-engine: engine for fine-tuning and serving LLMs [5]\n - Lamini: hosted? closed source? solution for training LLMs [6]\n - GPT4All: free, locally run chatbot [7]\n - SkyPilot: framework for running LLMs, AI, and batch jobs [8]\n - HuggingFace Transformers: APIs and tools to download and train models (via Pytorch) [9]\n - RAGStack: Deploy a private ChatGPT alternative hosted within your VPC [14]\n - Simon's `llm` tool [15]\n - AutoGPTQ: GPTQ algorithm based quantization package [10]\n - QLoRA: finetuning of quantized LLMs [11]\n - bitsandbytes: CUDA functions for PyTorch [12]\n - SkyPilot QLoRA [13]\n - https://www.youtube.com/watch?v=eeM6V5aPjhk\n - https://www.youtube.com/watch?v=TYgtG2Th6fI (jawerty's example)\n - [1] https://github.com/vllm-project/vllm\n - [2] https://github.com/jmorganca/ollama\n - [3] https://github.com/ggerganov/llama.cpp\n - [4] https://github.com/abetlen/llama-cpp-python\n - [5] https://github.com/scaleapi/llm-engine\n - [6] https://www.lamini.ai/\n - [7] https://github.com/nomic-ai/gpt4all\n - [8] https://github.com/skypilot-org/skypilot\n - [9] https://github.com/huggingface/transformers/releases\n - [10] https://github.com/PanQiWei/AutoGPTQ\n - [11] https://github.com/artidoro/qlora\n - [12] https://github.com/TimDettmers/bitsandbytes\n - [13] https://github.com/artidoro/qlora/pull/132\n - [14] https://github.com/psychic-api/rag-stack\n - [15] https://simonwillison.net/2023/Aug/1/llama-2-mac/Engines/APIs:
- vllm: Inference and serving engine for LLMs (none quantized models only?) [1]\n - ollama: Go project to run, create and share LLMs [2]\n - llama.cpp: Inference of LLaMA models in C/C++ w/UI (including quantized models) [3]\n - llama-cpp-python: run OpenAI compatible API bindings for llama.cpp [4]\n - llm-engine: engine for fine-tuning and serving LLMs [5]\n - Lamini: hosted? closed source? solution for training LLMs [6]\n - GPT4All: free, locally run chatbot [7]\n - SkyPilot: framework for running LLMs, AI, and batch jobs [8]\n - HuggingFace Transformers: APIs and tools to download and train models (via Pytorch) [9]\n - RAGStack: Deploy a private ChatGPT alternative hosted within your VPC [14]\n - Simon's `llm` tool [15]\n - AutoGPTQ: GPTQ algorithm based quantization package [10]\n - QLoRA: finetuning of quantized LLMs [11]\n - bitsandbytes: CUDA functions for PyTorch [12]\n - SkyPilot QLoRA [13]\n - https://www.youtube.com/watch?v=eeM6V5aPjhk\n - https://www.youtube.com/watch?v=TYgtG2Th6fI (jawerty's example)\n - [1] https://github.com/vllm-project/vllm\n - [2] https://github.com/jmorganca/ollama\n - [3] https://github.com/ggerganov/llama.cpp\n - [4] https://github.com/abetlen/llama-cpp-python\n - [5] https://github.com/scaleapi/llm-engine\n - [6] https://www.lamini.ai/\n - [7] https://github.com/nomic-ai/gpt4all\n - [8] https://github.com/skypilot-org/skypilot\n - [9] https://github.com/huggingface/transformers/releases\n - [10] https://github.com/PanQiWei/AutoGPTQ\n - [11] https://github.com/artidoro/qlora\n - [12] https://github.com/TimDettmers/bitsandbytes\n - [13] https://github.com/artidoro/qlora/pull/132\n - [14] https://github.com/psychic-api/rag-stack\n - [15] https://simonwillison.net/2023/Aug/1/llama-2-mac/All without Ray or TRL or DeepSpeed etc., workers talk to the trainer over stdlib HTTP."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: NanoRL \u2013 RL training for LLMs in ~1,800 lines"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/alex000kim/nanoRL"}},"_tags":["story","author_alex000kim","story_49286216","show_hn"],"author":"alex000kim","children":[49406866],"created_at":"2026-08-13T14:08:54Z","created_at_i":1786630134,"num_comments":0,"objectID":"49286216","points":11,"story_id":49286216,"story_text":"The smallest async RL trainer I could write: one loop that runs REINFORCE on CartPole on a laptop and async GRPO on a cluster \n(e.g. 8xH100 trainer, 8 vLLM workers, ran as a [SkyPilot job group](https://docs.skypilot.ai/en/latest/examples/job-groups.html) on k8s ).
All without Ray or TRL or DeepSpeed etc., workers talk to the trainer over stdlib HTTP.","title":"Show HN: NanoRL \u2013 RL training for LLMs in ~1,800 lines","updated_at":"2026-08-23T07:51:20Z","url":"https://github.com/alex000kim/nanoRL"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"coloneltambo"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"I'm thinking something that would let me provision and shut down gpu instances across all the clouds with GPUs through one api interface. I know skypilot exists but that's more from training than just infra."},"title":{"matchLevel":"none","matchedWords":[],"value":"A unified API for all the GPU clouds?"}},"_tags":["story","author_coloneltambo","story_36147120","ask_hn"],"author":"coloneltambo","children":[36148022],"created_at":"2023-06-01T02:57:56Z","created_at_i":1685588276,"num_comments":3,"objectID":"36147120","points":4,"story_id":36147120,"story_text":"I'm thinking something that would let me provision and shut down gpu instances across all the clouds with GPUs through one api interface. I know skypilot exists but that's more from training than just infra.","title":"A unified API for all the GPU clouds?","updated_at":"2024-12-24T10:24:33Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"We made a simple guide on how to run LLaMA chatbots on any cloud(s) of your choosing using one CLI command, powered by the open-source SkyPilot project.
Highlights\n- cost reduction\n- better GPU availability\n- cloud portability (big three + Lambda supported)"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Run LLaMA LLM chatbots on any cloud with one click"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://github.com/skypilot-org/skypilot/tree/master/examples/llama-llm-chatbots"}},"_tags":["story","author_covi","story_35256363","show_hn"],"author":"covi","children":[35276865],"created_at":"2023-03-22T03:14:27Z","created_at_i":1679454867,"num_comments":1,"objectID":"35256363","points":4,"story_id":35256363,"story_text":"We made a simple guide on how to run LLaMA chatbots on any cloud(s) of your choosing using one CLI command, powered by the open-source SkyPilot project.
Highlights\n- cost reduction\n- better GPU availability\n- cloud portability (big three + Lambda supported)","title":"Show HN: Run LLaMA LLM chatbots on any cloud with one click","updated_at":"2024-09-20T13:39:09Z","url":"https://github.com/skypilot-org/skypilot/tree/master/examples/llama-llm-chatbots"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"cuda-oom"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"I've been looking at job orchestration tools and there's so many ways to submit a training job.\nAre you using Flyte? SkyPilot? Does your infra team provide you with a Slurm cluster?\nWhat tools/frameworks support GPUs out of the box?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What are you using to submit ML training jobs?"}},"_tags":["story","author_cuda-oom","story_44672043","ask_hn"],"author":"cuda-oom","created_at":"2025-07-24T15:38:52Z","created_at_i":1753371532,"num_comments":0,"objectID":"44672043","points":2,"story_id":44672043,"story_text":"I've been looking at job orchestration tools and there's so many ways to submit a training job.\nAre you using Flyte? SkyPilot? Does your infra team provide you with a Slurm cluster?\nWhat tools/frameworks support GPUs out of the box?","title":"Ask HN: What are you using to submit ML training jobs?","updated_at":"2025-07-24T17:35:05Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hopechong"},"title":{"matchLevel":"none","matchedWords":[],"value":"Scaling Karpathy's Autoresearch: What Happens When the Agent Gets a GPU Cluster"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/scaling-autoresearch/"}},"_tags":["story","author_hopechong","story_47442435"],"author":"hopechong","children":[47443041,47443340,47443341,47443593,47443710,47444220,47444889,47445228,47445281,47445312,47445606,47446165,47446219,47446291,47446922,47446999,47447913,47448925,47449152,47450048,47450475,47450963,47451589,47451879,47452025,47452534,47452554,47452586,47452646,47452650,47454623,47454729,47459143,47459513,47592726],"created_at":"2026-03-19T16:55:45Z","created_at_i":1773939345,"num_comments":94,"objectID":"47442435","points":237,"story_id":47442435,"title":"Scaling Karpathy's Autoresearch: What Happens When the Agent Gets a GPU Cluster","updated_at":"2026-05-09T08:03:05Z","url":"https://blog.skypilot.co/scaling-autoresearch/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hopechong"},"title":{"matchLevel":"none","matchedWords":[],"value":"Research-Driven Agents: When an agent reads before it codes"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/research-driven-agents/"}},"_tags":["story","author_hopechong","story_47706141"],"author":"hopechong","children":[47706142,47707538,47707905,47707952,47708049,47708143,47708318,47708430,47708489,47708582,47708668,47708734,47708830,47709148,47709721,47709749,47709900,47710034,47710166,47710647,47710723,47711088,47711151,47711505,47711646,47711706,47711798,47712343,47713094,47713368,47713854,47714257,47714318,47715514,47715807,47716707,47717405,47718480,47719535,47723611,47724125],"created_at":"2026-04-09T16:58:49Z","created_at_i":1775753929,"num_comments":54,"objectID":"47706141","points":214,"story_id":47706141,"title":"Research-Driven Agents: When an agent reads before it codes","updated_at":"2026-07-28T18:50:06Z","url":"https://blog.skypilot.co/research-driven-agents/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"Cookbook: Finetuning Llama 2 in your own cloud environment, privately"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/finetuning-llama2-operational-guide/"}},"_tags":["story","author_covi","story_36975245"],"author":"covi","children":[36975496,36976041,36976741,36977912,36977927],"created_at":"2023-08-02T18:50:15Z","created_at_i":1691002215,"num_comments":13,"objectID":"36975245","points":119,"story_id":36975245,"title":"Cookbook: Finetuning Llama 2 in your own cloud environment, privately","updated_at":"2024-09-20T14:50:32Z","url":"https://blog.skypilot.co/finetuning-llama2-operational-guide/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hopechong"},"title":{"matchLevel":"none","matchedWords":[],"value":"Setting up OpenClaw on a cloud VM"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/openclaw-on-skypilot/"}},"_tags":["story","author_hopechong","story_47183503"],"author":"hopechong","children":[47183504,47183796,47183803,47183820,47183842,47183900,47183929,47184064,47184396,47184450,47184455,47184613,47184658,47184676,47184696,47184760,47184912,47185044,47185508,47185643,47187702,47191371,47195041,47266674],"created_at":"2026-02-27T18:06:26Z","created_at_i":1772215586,"num_comments":62,"objectID":"47183503","points":84,"story_id":47183503,"title":"Setting up OpenClaw on a cloud VM","updated_at":"2026-07-02T15:30:09Z","url":"https://blog.skypilot.co/openclaw-on-skypilot/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"Using DeepSeek R1 for RAG: Do's and Don'ts"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/deepseek-rag/"}},"_tags":["story","author_covi","story_43190614"],"author":"covi","children":[43190679,43190848],"created_at":"2025-02-27T02:19:27Z","created_at_i":1740622767,"num_comments":0,"objectID":"43190614","points":13,"story_id":43190614,"title":"Using DeepSeek R1 for RAG: Do's and Don'ts","updated_at":"2025-02-27T04:26:40Z","url":"https://blog.skypilot.co/deepseek-rag/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"Run LLaMA LLM chatbots on any cloud with one click"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/llama-llm-chatbots-on-any-cloud/"}},"_tags":["story","author_covi","story_35241732"],"author":"covi","created_at":"2023-03-21T02:14:11Z","created_at_i":1679364851,"num_comments":0,"objectID":"35241732","points":12,"story_id":35241732,"title":"Run LLaMA LLM chatbots on any cloud with one click","updated_at":"2024-09-20T13:37:27Z","url":"https://blog.skypilot.co/llama-llm-chatbots-on-any-cloud/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"alex000kim"},"title":{"matchLevel":"none","matchedWords":[],"value":"RL Is Bottlenecked by Inference. Scale It Independently"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://skypilot.ai/blog/slime-agentic-rl"}},"_tags":["story","author_alex000kim","story_49183059"],"author":"alex000kim","children":[49183060,49183587],"created_at":"2026-08-05T14:03:17Z","created_at_i":1785938597,"num_comments":2,"objectID":"49183059","points":11,"story_id":49183059,"title":"RL Is Bottlenecked by Inference. Scale It Independently","updated_at":"2026-08-06T16:01:38Z","url":"https://skypilot.ai/blog/slime-agentic-rl"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hopechong"},"title":{"matchLevel":"none","matchedWords":[],"value":"GPU Compass: Navigate the GPU Frontier Across 20 Clouds and 2K+ Offerings"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://gpus.skypilot.co/"}},"_tags":["story","author_hopechong","story_47866256"],"author":"hopechong","children":[47866257],"created_at":"2026-04-22T17:01:25Z","created_at_i":1776877285,"num_comments":1,"objectID":"47866256","points":11,"story_id":47866256,"title":"GPU Compass: Navigate the GPU Frontier Across 20 Clouds and 2K+ Offerings","updated_at":"2026-04-22T19:23:53Z","url":"https://gpus.skypilot.co/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Bogdanp"},"title":{"matchLevel":"none","matchedWords":[],"value":"Network and Storage Benchmarks for LLM Training on the Cloud"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://maknee.github.io/blog/2025/Network-And-Storage-Training-Skypilot/"}},"_tags":["story","author_Bogdanp","story_45212260"],"author":"Bogdanp","created_at":"2025-09-11T14:42:10Z","created_at_i":1757601730,"num_comments":0,"objectID":"45212260","points":10,"story_id":45212260,"title":"Network and Storage Benchmarks for LLM Training on the Cloud","updated_at":"2026-03-05T22:39:04Z","url":"https://maknee.github.io/blog/2025/Network-And-Storage-Training-Skypilot/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rombr"},"title":{"matchLevel":"none","matchedWords":[],"value":"Migrating from Slurm to Kubernetes"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/slurm-to-k8s-migration/"}},"_tags":["story","author_rombr","story_46978421"],"author":"rombr","children":[46978678,46979771],"created_at":"2026-02-11T18:00:59Z","created_at_i":1770832859,"num_comments":5,"objectID":"46978421","points":9,"story_id":46978421,"title":"Migrating from Slurm to Kubernetes","updated_at":"2026-03-05T23:34:54Z","url":"https://blog.skypilot.co/slurm-to-k8s-migration/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"zhwu"},"title":{"matchLevel":"none","matchedWords":[],"value":"Guide on fine-tuning your own Vicuna on Llama-2"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://twitter.com/skypilot_org/status/1686809528069988352"}},"_tags":["story","author_zhwu","story_36989623"],"author":"zhwu","created_at":"2023-08-03T18:18:03Z","created_at_i":1691086683,"num_comments":0,"objectID":"36989623","points":9,"story_id":36989623,"title":"Guide on fine-tuning your own Vicuna on Llama-2","updated_at":"2024-09-20T14:51:57Z","url":"https://twitter.com/skypilot_org/status/1686809528069988352"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rombr"},"title":{"matchLevel":"none","matchedWords":[],"value":"Slurm vs. K8s for AI Infra: Academic HPC vs. Cloud-Native Reality"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/slurm-vs-k8s/"}},"_tags":["story","author_rombr","story_44759327"],"author":"rombr","created_at":"2025-08-01T16:51:40Z","created_at_i":1754067100,"num_comments":0,"objectID":"44759327","points":8,"story_id":44759327,"title":"Slurm vs. K8s for AI Infra: Academic HPC vs. Cloud-Native Reality","updated_at":"2025-08-18T13:01:24Z","url":"https://blog.skypilot.co/slurm-vs-k8s/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gaocegege"},"title":{"matchLevel":"none","matchedWords":[],"value":"Should AI workloads run on Kubernetes?"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/ai-on-kubernetes/"}},"_tags":["story","author_gaocegege","story_41066340"],"author":"gaocegege","children":[41067331],"created_at":"2024-07-25T08:47:04Z","created_at_i":1721897224,"num_comments":1,"objectID":"41066340","points":7,"story_id":41066340,"title":"Should AI workloads run on Kubernetes?","updated_at":"2024-09-20T17:30:26Z","url":"https://blog.skypilot.co/ai-on-kubernetes/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rombr"},"title":{"matchLevel":"none","matchedWords":[],"value":"Chat with your PDFs \u2013 Self-hosted LocalGPT on any cloud"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://github.com/skypilot-org/skypilot/blob/master/llm/localgpt/README.md"}},"_tags":["story","author_rombr","story_36633657"],"author":"rombr","created_at":"2023-07-07T15:57:25Z","created_at_i":1688745445,"num_comments":0,"objectID":"36633657","points":6,"story_id":36633657,"title":"Chat with your PDFs \u2013 Self-hosted LocalGPT on any cloud","updated_at":"2024-09-20T14:34:05Z","url":"https://github.com/skypilot-org/skypilot/blob/master/llm/localgpt/README.md"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"RL Doesn't Work on Slurm"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/rl-doesnt-work-on-slurm/"}},"_tags":["story","author_covi","story_48238565"],"author":"covi","created_at":"2026-05-22T17:04:33Z","created_at_i":1779469473,"num_comments":0,"objectID":"48238565","points":5,"story_id":48238565,"title":"RL Doesn't Work on Slurm","updated_at":"2026-05-22T20:11:12Z","url":"https://blog.skypilot.co/rl-doesnt-work-on-slurm/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"emschwartz"},"title":{"matchLevel":"none","matchedWords":[],"value":"Abusing SQLite to Handle Concurrency"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/abusing-sqlite-to-handle-concurrency/"}},"_tags":["story","author_emschwartz","story_43258658"],"author":"emschwartz","created_at":"2025-03-04T19:00:10Z","created_at_i":1741114810,"num_comments":0,"objectID":"43258658","points":5,"story_id":43258658,"title":"Abusing SQLite to Handle Concurrency","updated_at":"2025-03-05T02:07:01Z","url":"https://blog.skypilot.co/abusing-sqlite-to-handle-concurrency/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"Guide: Finetune Llama 3.1 on your infra"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://github.com/skypilot-org/skypilot/tree/master/llm/llama-3_1-finetuning"}},"_tags":["story","author_covi","story_41053169"],"author":"covi","created_at":"2024-07-24T02:43:37Z","created_at_i":1721789017,"num_comments":0,"objectID":"41053169","points":5,"story_id":41053169,"title":"Guide: Finetune Llama 3.1 on your infra","updated_at":"2024-09-20T17:29:01Z","url":"https://github.com/skypilot-org/skypilot/tree/master/llm/llama-3_1-finetuning"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"SkyServe: 50% Cheaper AI Serving on Any Cloud with High Availability"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/introducing-sky-serve/"}},"_tags":["story","author_covi","story_39449907"],"author":"covi","created_at":"2024-02-21T03:43:55Z","created_at_i":1708487035,"num_comments":0,"objectID":"39449907","points":5,"story_id":39449907,"title":"SkyServe: 50% Cheaper AI Serving on Any Cloud with High Availability","updated_at":"2024-09-20T16:26:49Z","url":"https://blog.skypilot.co/introducing-sky-serve/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"covi"},"title":{"matchLevel":"none","matchedWords":[],"value":"Covariant: Moving from on-prem to the cloud for 4x faster AI development"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/covariant/"}},"_tags":["story","author_covi","story_37661761"],"author":"covi","children":[37662128],"created_at":"2023-09-26T16:30:29Z","created_at_i":1695745829,"num_comments":1,"objectID":"37661761","points":4,"story_id":37661761,"title":"Covariant: Moving from on-prem to the cloud for 4x faster AI development","updated_at":"2024-09-20T15:14:22Z","url":"https://blog.skypilot.co/covariant/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rombr"},"title":{"matchLevel":"none","matchedWords":[],"value":"AI on Kubernetes Without the Pain"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["skypilot"],"value":"https://blog.skypilot.co/ai-on-kubernetes/"}},"_tags":["story","author_rombr","story_42678920"],"author":"rombr","created_at":"2025-01-13T01:13:23Z","created_at_i":1736730803,"num_comments":0,"objectID":"42678920","points":4,"story_id":42678920,"title":"AI on Kubernetes Without the Pain","updated_at":"2025-01-13T02:07:47Z","url":"https://blog.skypilot.co/ai-on-kubernetes/"}],"hitsPerPage":50,"nbHits":83,"nbPages":2,"page":0,"params":"query=SkyPilot&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":11,"processingTimingsMS":{"_request":{"roundTrip":17},"afterFetch":{"format":{"total":1},"merge":{"mergeLoop":{"prepareNextHit":3,"total":3},"total":3},"total":3},"fetch":{"query":6,"total":7},"total":11},"query":"SkyPilot","serverTimeMS":13}