{
    "version": "https://jsonfeed.org/version/1",
    "title": "twango.dev",
    "home_page_url": "https://twango.dev/writing",
    "feed_url": "https://twango.dev/feed.json",
    "description": "Writing from twango.dev",
    "author": {
        "name": "James Ding",
        "url": "https://twango.dev"
    },
    "items": [
        {
            "id": "https://twango.dev/writing/jetplay",
            "content_html": "<p>A plugin to allow playback of audio and video inside JetBrains IDEs (IntelliJ, PyCharm, WebStorm, and the rest), right in the editor tab.</p>\n<p><a href=\"https://twango.dev/writing/jetplay\">Read on twango.dev →</a></p>",
            "url": "https://twango.dev/writing/jetplay",
            "title": "Jetplay: Audio and Video Playback for JetBrains IDEs",
            "summary": "A plugin to allow playback of audio and video inside JetBrains IDEs (IntelliJ, PyCharm, WebStorm, and the rest), right in the editor tab.",
            "date_modified": "2026-06-16T00:00:00.000Z",
            "date_published": "2026-06-16T00:00:00.000Z",
            "author": {
                "name": "James Ding",
                "url": "https://twango.dev"
            },
            "tags": [
                "engineering",
                "jetbrains",
                "media",
                "plugin"
            ]
        },
        {
            "id": "https://twango.dev/writing/tartanaviation-atc-labels",
            "content_html": "<p>Three Hugging Face datasets from CMU's TartanAviation air traffic control corpus: paired audio and ADS-B clips, a half-million-utterance VAD split, and a full ASR label set. Why decode-time biasing and LLM correction lost to a plain ROVER vote, and the one signal that moved label quality.</p>\n<p><a href=\"https://twango.dev/writing/tartanaviation-atc-labels\">Read on twango.dev →</a></p>",
            "url": "https://twango.dev/writing/tartanaviation-atc-labels",
            "title": "The TartanAviation ATC Collection: Audio, ADS-B, and 531k Labels",
            "summary": "Three Hugging Face datasets from CMU's TartanAviation air traffic control corpus: paired audio and ADS-B clips, a half-million-utterance VAD split, and a full ASR label set. Why decode-time biasing and LLM correction lost to a plain ROVER vote, and the one signal that moved label quality.",
            "date_modified": "2026-06-05T00:00:00.000Z",
            "date_published": "2026-06-05T00:00:00.000Z",
            "author": {
                "name": "James Ding",
                "url": "https://twango.dev"
            },
            "tags": [
                "engineering",
                "ml",
                "asr",
                "atc",
                "dataset"
            ]
        },
        {
            "id": "https://twango.dev/writing/gemma4-asr-benchmark",
            "content_html": "<p>Benchmarking the Gemma 4 family (E2B, E4B, and the new 12B) across all eight Open ASR Leaderboard short-form datasets, batched. WER, RTFx, where a general-purpose multimodal LLM holds up, and where it doesn't.</p>\n<p><a href=\"https://twango.dev/writing/gemma4-asr-benchmark\">Read on twango.dev →</a></p>",
            "url": "https://twango.dev/writing/gemma4-asr-benchmark",
            "title": "How Good is Gemma 4 at ASR Tasks? Benchmarking Against the Open ASR Leaderboard",
            "summary": "Benchmarking the Gemma 4 family (E2B, E4B, and the new 12B) across all eight Open ASR Leaderboard short-form datasets, batched. WER, RTFx, where a general-purpose multimodal LLM holds up, and where it doesn't.",
            "date_modified": "2026-06-04T00:00:00.000Z",
            "date_published": "2026-06-04T00:00:00.000Z",
            "author": {
                "name": "James Ding",
                "url": "https://twango.dev"
            },
            "tags": [
                "engineering",
                "ml",
                "benchmark"
            ]
        }
    ]
}