[Enhancement] Add README quickstart examples for all Deepgram product APIs

[deepgram-robot]

Summary

Add comprehensive quickstart code examples to the Java SDK README covering all major Deepgram APIs: pre-recorded transcription, live streaming transcription, text-to-speech, audio intelligence, and voice agents.

Problem it solves

The Java SDK was just launched (v0.1.0, 2026-03-27) and developers evaluating it need immediate, copy-pasteable examples to understand the API surface. Java developers are particularly sensitive to boilerplate clarity — they need to see the builder patterns, response types, and async handling before committing to a dependency. With competing Java STT SDKs recently discontinued, this is a unique window to capture Java developers with excellent onboarding documentation.

Proposed API

README should include at minimum:

// Pre-recorded transcription
DeepgramClient client = new DeepgramClient("API_KEY");
PreRecordedResponse response = client.listen().preRecorded()
    .url("https://example.com/audio.wav")
    .model("nova-3")
    .smartFormat(true)
    .transcribe();
System.out.println(response.getResults().getChannels().get(0).getAlternatives().get(0).getTranscript());

// Live streaming
LiveTranscriptionClient live = client.listen().live()
    .model("nova-3")
    .onTranscript(result -> System.out.println(result.getTranscript()))
    .connect();

// Text-to-speech
byte[] audio = client.speak()
    .text("Hello from Deepgram")
    .model("aura-asteria-en")
    .generate();
Files.write(Path.of("output.wav"), audio);

Acceptance criteria

• README includes quickstart examples for: pre-recorded STT, live STT, TTS, and at least one Intelligence feature
• Each example is self-contained and copy-pasteable
• Examples show both sync and async patterns where available
• Documented with usage example
• Compatible with existing API
• Maven/Gradle dependency snippet included

---

Raised by the DX intelligence system.