Improve interview coach live flow

Co-authored-by: Codex <codex@openai.com>

README.md

@@ -12,6 +12,8 @@ pinned: false
 
 Interview Coach is a local-first assistant for live technical interviews. It listens to noisy interview audio, extracts the actual Data Science, ML, AI, or System Design question, and shows a compact coaching card with important pointers while the candidate is answering.
 
+Development note: the live transcription flow, coaching-card timing, and transcript extraction loop were iterated with Codex assistance.
+
 ## Why It Matters
 
 The goal is not to handhold the candidate through the interview or generate a scripted answer. The goal is to give timely, high-signal reminders so the candidate can cover the important parts of their own answer naturally.

app.py

@@ -12,6 +12,7 @@ import numpy as np
 from agents.evaluator import EvaluationAgent
 from agents.hf_chat import HuggingFaceChatModel
 from agents.topic_pattern import TopicPatternAgent
+<<<<<<< HEAD
 from config import (
     APP_HOST,
     APP_PORT,
@@ -21,6 +22,9 @@ from config import (
     STREAMING_WHISPER_MODEL,
     TOPIC_PATTERN_MODEL,
 )
+=======
+from config import APP_HOST, APP_PORT, BASE_DIR, GENERAL_LLM_MODEL, HF_SPACE_MODE, STREAMING_WHISPER_MODEL
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 from db.queries import (
     add_evaluation,
     append_exchange_answer,
@@ -418,7 +422,11 @@ async def classify_topic_and_steps(question: str) -> dict[str, Any]:
             "model_unavailable": True,
             "message": (
                 "Topic/steps model unavailable. Expected Hugging Face model "
+<<<<<<< HEAD
                 f"{TOPIC_PATTERN_MODEL}.{suffix}"
+=======
+                f"vadirajkrishna/interview-coach-3b.{suffix}"
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
             ),
         }
 

article.md

@@ -39,7 +39,11 @@ The project uses a multi-model approach. Instead of forcing one model to handle
 | --- | ---: | --- |
 | `mlx-community/whisper-tiny` | ~39M parameters | Fast streaming transcription for live audio |
 | `mlx-community/whisper-small-mlx` | ~244M parameters | Higher-quality local transcription when needed |
+<<<<<<< HEAD
 | `build-small-hackathon/interview-coach-3b` | 3B base model plus LoRA adapter | Fine-tuned topic and pattern detection |
+=======
+| `vadirajkrishna/interview-coach-3b` | 3B base model plus LoRA adapter | Fine-tuned topic and pattern detection |
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 | `Qwen/Qwen2.5-3B-Instruct` | ~3B parameters | General reasoning, coaching hint generation, transcript cleanup, and evaluation |
 | SQLite | Local database | Session, exchange, and evaluation storage |
 

config.py

@@ -14,7 +14,11 @@ STT_BACKEND = os.environ.get("INTERVIEW_COACH_STT_BACKEND", "transformers" if HF
 HF_WHISPER_MODEL = os.environ.get("INTERVIEW_COACH_HF_WHISPER_MODEL", "openai/whisper-tiny")
 APP_HOST = os.environ.get("INTERVIEW_COACH_HOST", "0.0.0.0" if HF_SPACE_MODE else "127.0.0.1")
 APP_PORT = 7860
+<<<<<<< HEAD
 TOPIC_PATTERN_MODEL = os.environ.get("INTERVIEW_COACH_TOPIC_PATTERN_MODEL", "build-small-hackathon/interview-coach-3b")
+=======
+TOPIC_PATTERN_MODEL = os.environ.get("INTERVIEW_COACH_TOPIC_PATTERN_MODEL", "vadirajkrishna/interview-coach-3b")
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 TOPIC_PATTERN_BASE_MODEL = os.environ.get("INTERVIEW_COACH_TOPIC_PATTERN_BASE_MODEL", GENERAL_LLM_MODEL)
 USE_TOPIC_PATTERN_MODEL = os.environ.get("INTERVIEW_COACH_USE_TOPIC_PATTERN_MODEL", "1").strip().lower() not in {
     "0",

fine-tune/lora-finetune.py

@@ -41,7 +41,11 @@ import modal
 SCRIPT_DIR = Path(__file__).resolve().parent
 
 # Your Hugging Face username (https://huggingface.co/settings/profile)
+<<<<<<< HEAD
 HF_USERNAME = "build-small-hackathon"
+=======
+HF_USERNAME = "vadirajkrishna"
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 
 # Name for the fine-tuned model repo on Hugging Face
 HF_MODEL_NAME = "interview-coach-3b"

requirements.txt

@@ -1,3 +1,4 @@
+<<<<<<< HEAD
 accelerate>=1.9.0
 aiosqlite>=0.20.0
 gradio>=6.16.0
@@ -6,9 +7,22 @@ imageio-ffmpeg>=0.5.1
 langgraph>=0.2.0
 mlx-whisper>=0.4.0; platform_system == "Darwin"
 numpy>=1.26.0
+=======
+aiosqlite>=0.20.0
+accelerate>=1.9.0
+gradio>=4.44.0
+imageio-ffmpeg>=0.5.1
+langgraph>=0.2.0
+mlx-whisper>=0.4.0; platform_system == "Darwin"
+modal>=1.1.0
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 peft>=0.16.0
 pyyaml>=6.0.2
 scipy>=1.11.0
 sounddevice>=0.5.1; platform_system == "Darwin"
+<<<<<<< HEAD
 torch>=2.4.0
+=======
+torch>=2.3.0
+>>>>>>> c0f39ad (initial commit - InterviewCopilotLocal)
 transformers>=4.54.0