azure-ai-search-recruitment-agent.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 🚀 HR Resume Search using Assistant API File Search\n",
    "\n",
    "This notebook demonstrates how to use the OpenAI Assistant API's File Search capability\n",
    "to analyze resumes more effectively. The implementation follows best practices for\n",
    "file handling and annotation support.\n",
    "\n",
    "## Prerequisites\n",
    "\n",
    "- Azure OpenAI resource with Assistant API access\n",
    "- Semantic Kernel v1.16+\n",
    "- Python 3.8+\n",
    "\n",
    "## Setup\n",
    "\n",
    "First, let's install the required packages:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install semantic-kernel==1.16.0 python-dotenv aiofiles nest_asyncio azure-search-documents"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## Import Dependencies and Configure Environment\n",
    " \n",
    "Let's import all necessary libraries and set up our environment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "import asyncio\n",
    "import os\n",
    "from pathlib import Path\n",
    "from typing import List\n",
    "from dataclasses import dataclass\n",
    "\n",
    "from semantic_kernel.kernel import Kernel\n",
    "from semantic_kernel.agents.open_ai.azure_assistant_agent import AzureAssistantAgent\n",
    "from semantic_kernel.contents.chat_message_content import ChatMessageContent\n",
    "from semantic_kernel.contents.streaming_annotation_content import StreamingAnnotationContent\n",
    "from semantic_kernel.contents.utils.author_role import AuthorRole\n",
    "from dotenv import load_dotenv\n",
    "\n",
    "# Enable notebook async support\n",
    "import nest_asyncio\n",
    "nest_asyncio.apply()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Configuration Management\n",
    "\n",
    "We'll create a configuration class to manage our Azure OpenAI settings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from dataclasses import dataclass\n",
    "from dotenv import load_dotenv\n",
    "\n",
    "\n",
    "@dataclass\n",
    "class AzureConfig:\n",
    "    \"\"\"Configuration for Azure OpenAI Assistant.\"\"\"\n",
    "\n",
    "    api_key: str\n",
    "    endpoint: str\n",
    "    deployment_name: str\n",
    "    api_version: str = \"2024-10-01-preview\"\n",
    "\n",
    "    @classmethod\n",
    "    def from_env(cls) -> \"AzureConfig\":\n",
    "        \"\"\"Load configuration from environment variables.\"\"\"\n",
    "        load_dotenv()\n",
    "\n",
    "        # Get environment variables\n",
    "        api_key = os.getenv(\"AZURE_OPENAI_API_KEY\")\n",
    "        endpoint = os.getenv(\"AZURE_OPENAI_ENDPOINT\")\n",
    "        deployment_name = os.getenv(\"AZURE_OPENAI_CHAT_COMPLETION_DEPLOYED_MODEL_NAME\")\n",
    "\n",
    "        if not all([api_key, endpoint, deployment_name]):\n",
    "            missing = [\n",
    "                var\n",
    "                for var, val in {\n",
    "                    \"AZURE_OPENAI_API_KEY\": api_key,\n",
    "                    \"AZURE_OPENAI_ENDPOINT\": endpoint,\n",
    "                    \"AZURE_OPENAI_CHAT_COMPLETION_DEPLOYED_MODEL_NAME\": deployment_name,\n",
    "                }.items()\n",
    "                if not val\n",
    "            ]\n",
    "            raise ValueError(\n",
    "                f\"Missing required environment variables: {', '.join(missing)}\"\n",
    "            )\n",
    "\n",
    "        # Clean up endpoint URL\n",
    "        if endpoint:\n",
    "            endpoint = endpoint.rstrip(\"/\")\n",
    "\n",
    "        return cls(\n",
    "            api_key=api_key,\n",
    "            endpoint=endpoint,\n",
    "            deployment_name=deployment_name,\n",
    "            api_version=\"2024-10-01-preview\",\n",
    "        )\n",
    "\n",
    "    \"\"\"Configuration for Azure OpenAI Assistant.\"\"\"\n",
    "    api_key: str\n",
    "    endpoint: str\n",
    "    deployment_name: str\n",
    "    api_version: str = \"2024-10-01-preview\"\n",
    "\n",
    "    @classmethod\n",
    "    def from_env(cls) -> \"AzureConfig\":\n",
    "        \"\"\"Load configuration from environment variables.\"\"\"\n",
    "        load_dotenv()\n",
    "\n",
    "        # Get environment variables\n",
    "        api_key = os.getenv(\"AZURE_OPENAI_API_KEY\")\n",
    "        endpoint = os.getenv(\"AZURE_OPENAI_ENDPOINT\")\n",
    "        deployment_name = os.getenv(\"AZURE_OPENAI_CHAT_COMPLETION_DEPLOYED_MODEL_NAME\")\n",
    "        api_version = os.getenv(\"AZURE_OPENAI_API_VERSION\", \"2024-10-01-preview\")\n",
    "\n",
    "        # Validate required variables\n",
    "        if not all([api_key, endpoint, deployment_name]):\n",
    "            missing = [\n",
    "                var\n",
    "                for var, val in {\n",
    "                    \"AZURE_OPENAI_API_KEY\": api_key,\n",
    "                    \"AZURE_OPENAI_ENDPOINT\": endpoint,\n",
    "                    \"AZURE_OPENAI_CHAT_COMPLETION_DEPLOYED_MODEL_NAME\": deployment_name,\n",
    "                }.items()\n",
    "                if not val\n",
    "            ]\n",
    "            raise ValueError(\n",
    "                f\"Missing required environment variables: {', '.join(missing)}\"\n",
    "            )\n",
    "\n",
    "        return cls(\n",
    "            api_key=api_key,\n",
    "            endpoint=endpoint,\n",
    "            deployment_name=deployment_name,\n",
    "            api_version=api_version,\n",
    "        )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## File Management\n",
    " \n",
    " Set up utilities to manage resume files and their paths."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_filepath_for_filename(filename: str) -> str:\n",
    "    \"\"\"Get the full path for a given filename.\"\"\"\n",
    "    base_directory = Path.cwd() / \"resumes\"\n",
    "    base_directory.mkdir(exist_ok=True)\n",
    "    return str(base_directory / filename)\n",
    "\n",
    "\n",
    "# Sample resume data\n",
    "sample_resumes = {\n",
    "    \"john_doe.txt\": \"\"\"\n",
    "    John Doe\n",
    "    Senior Software Engineer\n",
    "\n",
    "    Experience:\n",
    "    - Lead Developer at TechCorp (2019-Present)\n",
    "      * Led team of 5 developers on cloud migration project\n",
    "      * Implemented MLOps pipeline reducing deployment time by 60%\n",
    "      * Mentored junior developers and conducted code reviews\n",
    "    \n",
    "    - Senior Software Engineer at InnovSoft (2016-2019)\n",
    "      * Developed machine learning models for predictive maintenance\n",
    "      * Architected microservices infrastructure using Kubernetes\n",
    "      * Improved system performance by 40%\n",
    "\n",
    "    Skills:\n",
    "    - Programming: Python, Java, Go\n",
    "    - Cloud & DevOps: Kubernetes, Docker, AWS\n",
    "    - Machine Learning: TensorFlow, PyTorch, MLOps\n",
    "    - Leadership: Team Management, Mentoring\n",
    "    \n",
    "    Education:\n",
    "    - M.S. Computer Science, Tech University (2016)\n",
    "    - B.S. Computer Science, State University (2014)\n",
    "    \"\"\",\n",
    "    \n",
    "    \"jane_smith.txt\": \"\"\"\n",
    "    Jane Smith\n",
    "    AI Research Engineer\n",
    "\n",
    "    Experience:\n",
    "    - AI Research Lead at DataMinds (2020-Present)\n",
    "      * Published 3 papers on NLP architectures\n",
    "      * Developed novel attention mechanism improving accuracy by 25%\n",
    "      * Led research team of 3 PhD candidates\n",
    "    \n",
    "    - ML Engineer at AITech (2018-2020)\n",
    "      * Implemented computer vision models for autonomous systems\n",
    "      * Reduced model inference time by 35%\n",
    "      * Collaborated with cross-functional teams\n",
    "\n",
    "    Skills:\n",
    "    - Deep Learning: PyTorch, TensorFlow\n",
    "    - NLP: Transformers, BERT, GPT\n",
    "    - Research: Paper Writing, Experimentation\n",
    "    - Languages: Python, C++\n",
    "    \n",
    "    Education:\n",
    "    - PhD in Machine Learning, Tech Institute (2020)\n",
    "    - M.S. AI, Data University (2018)\n",
    "    \"\"\"\n",
    "}\n",
    "\n",
    "\n",
    "# Save resumes to files\n",
    "resume_files = []\n",
    "for filename, content in sample_resumes.items():\n",
    "    filepath = get_filepath_for_filename(filename)\n",
    "    Path(filepath).write_text(content, encoding=\"utf-8\")\n",
    "    resume_files.append(filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## Initialize Assistant\n",
    " \n",
    " Create an Azure Assistant Agent with file search capability enabled."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "async def create_hr_assistant() -> AzureAssistantAgent:\n",
    "    \"\"\"Create and configure the HR Assistant.\"\"\"\n",
    "    try:\n",
    "        config = AzureConfig.from_env()\n",
    "        kernel = Kernel()\n",
    "\n",
    "        print(f\"Initializing Assistant with:\")\n",
    "        print(f\"- Deployment: {config.deployment_name}\")\n",
    "        print(f\"- Endpoint: {config.endpoint}\")\n",
    "        print(f\"- API Version: {config.api_version}\")\n",
    "\n",
    "        # Get file paths for resumes\n",
    "        resume_paths = [get_filepath_for_filename(f) for f in resume_files]\n",
    "        print(f\"- Resume files: {resume_paths}\")\n",
    "\n",
    "        # Create the assistant with required configuration\n",
    "        agent = await AzureAssistantAgent.create(\n",
    "            kernel=kernel,\n",
    "            deployment_name=config.deployment_name,\n",
    "            endpoint=config.endpoint,\n",
    "            api_key=config.api_key,\n",
    "            api_version=config.api_version,\n",
    "            name=\"HR_Resume_Analyzer\",\n",
    "            instructions=\"\"\"\n",
    "            You are an expert HR assistant specialized in analyzing resumes and providing \n",
    "            detailed candidate evaluations.\n",
    "            \n",
    "            Guidelines:\n",
    "            1. Always analyze the resumes in the document store for your answers\n",
    "            2. Provide specific evidence and quotes from the resumes\n",
    "            3. Format responses using markdown for better readability\n",
    "            4. Compare candidates objectively based on their documented experience\n",
    "            5. Highlight quantifiable achievements and metrics\n",
    "            6. Include relevant education and certification details\n",
    "            \"\"\",\n",
    "            enable_file_search=True,\n",
    "            vector_store_filenames=resume_paths,\n",
    "            ai_model_id=config.deployment_name,  # Required parameter\n",
    "            metadata={\n",
    "                \"type\": \"hr_assistant\",\n",
    "                \"version\": \"1.0\",\n",
    "                \"capabilities\": \"resume_analysis,candidate_comparison\",\n",
    "            },\n",
    "            temperature=0.7,\n",
    "            top_p=0.95,\n",
    "        )\n",
    "\n",
    "        return agent\n",
    "\n",
    "    except Exception as e:\n",
    "        print(f\"Error creating assistant: {str(e)}\")\n",
    "        print(f\"Full exception details: {type(e).__name__}: {str(e)}\")\n",
    "        raise"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ## Query Interface\n",
    " \n",
    " Create an interface to interact with the assistant and handle responses with citations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "async def analyze_resumes():\n",
    "    \"\"\"Main function to interact with the HR Assistant.\"\"\"\n",
    "    print(\"Initializing HR Assistant...\")\n",
    "    agent = await create_hr_assistant()\n",
    "    \n",
    "    print(\"Creating conversation thread...\")\n",
    "    thread_id = await agent.create_thread()\n",
    "    \n",
    "    try:\n",
    "        while True:\n",
    "            user_input = input(\"\\nEnter your question (or 'exit' to quit): \")\n",
    "            if not user_input or user_input.lower() == \"exit\":\n",
    "                break\n",
    "            \n",
    "            await agent.add_chat_message(\n",
    "                thread_id=thread_id,\n",
    "                message=ChatMessageContent(\n",
    "                    role=AuthorRole.USER,\n",
    "                    content=user_input\n",
    "                )\n",
    "            )\n",
    "            \n",
    "            print(\"\\nAnalyzing resumes...\\n\")\n",
    "            footnotes: List[StreamingAnnotationContent] = []\n",
    "            \n",
    "            async for response in agent.invoke_stream(thread_id=thread_id):\n",
    "                footnotes.extend([\n",
    "                    item for item in response.items \n",
    "                    if isinstance(item, StreamingAnnotationContent)\n",
    "                ])\n",
    "                print(response.content, end=\"\", flush=True)\n",
    "            \n",
    "            if footnotes:\n",
    "                print(\"\\n\\nCitations:\")\n",
    "                for note in footnotes:\n",
    "                    print(f\"\\n• From {note.file_id}:\")\n",
    "                    print(f'  \"{note.quote}\"')\n",
    "    \n",
    "    finally:\n",
    "        print(\"\\nCleaning up resources...\")\n",
    "        if agent:\n",
    "            for file_id in agent.file_search_file_ids:\n",
    "                await agent.delete_file(file_id)\n",
    "            await agent.delete_thread(thread_id)\n",
    "            await agent.delete()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run the Analysis\n",
    "\n",
    "Execute the resume analysis interface to evaluate candidates. Each cell can be run multiple times for new analysis sessions.\n",
    "\n",
    "### Key Demo Questions:\n",
    "1. \"Create a technical competency matrix for both candidates focusing on: AI/ML expertise, cloud infrastructure, and leadership. Rate as Basic/Intermediate/Advanced with evidence.\"\n",
    "2. \"What are the most recent and relevant projects each candidate has led? Include team sizes and outcomes.\"\n",
    "3. \"Compare both candidates' experience with ML systems deployment and provide evidence of successful implementations.\"\n",
    "4. \"Create a final recommendation table with: top 3 strengths, growth areas, and risk factors for each candidate.\"\n",
    "\n",
    "# Additional Questions by Category (for reference):\n",
    "'''\n",
    "Initial Screening:\n",
    "- \"Summarize each candidate's minimum requirements compliance and flag any gaps.\"\n",
    "- \"Extract all quantifiable achievements (percentages, team sizes, metrics) from both resumes.\"\n",
    "- \"Compare their educational backgrounds and relevant certifications.\"\n",
    "\n",
    "Technical Deep-Dive:\n",
    "- \"What evidence exists of performance optimization? Include specific metrics.\"\n",
    "- \"Detail their experience with modern AI/ML tools and frameworks.\"\n",
    "- \"Compare their cloud and infrastructure experience with specific examples.\"\n",
    "\n",
    "Leadership Assessment:\n",
    "- \"Analyze their progression into leadership roles and project complexities.\"\n",
    "- \"What examples show stakeholder management and cross-functional collaboration?\"\n",
    "- \"Compare their mentoring experience and team development outcomes.\"\n",
    "\n",
    "Final Evaluation:\n",
    "- \"Calculate a job fit score (1-10) across technical skills, leadership, and innovation.\"\n",
    "- \"Which candidate shows stronger evidence of scaling AI systems in production?\"\n",
    "- \"Based on the role requirements, who would you shortlist? Provide supporting evidence.\"\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initializing HR Assistant...\n",
      "Initializing Assistant with:\n",
      "- Deployment: gpt-4o-mini\n",
      "- Endpoint: https://fsunavala-openai-eus.openai.azure.com/\n",
      "- API Version: 2024-10-01-preview\n",
      "- Resume files: ['c:\\\\Dev\\\\azure-ai-search-python-playground\\\\resumes\\\\john_doe.txt', 'c:\\\\Dev\\\\azure-ai-search-python-playground\\\\resumes\\\\jane_smith.txt']\n",
      "Creating conversation thread...\n",
      "\n",
      "Analyzing resumes...\n",
      "\n",
      "Here's a technical competency matrix for both candidates focusing on their expertise in AI/ML, cloud infrastructure, and leadership. Each competency is rated as Basic, Intermediate, or Advanced based on the evidence from their resumes.\n",
      "\n",
      "### Technical Competency Matrix\n",
      "\n",
      "| Competency          | Jane Smith                | John Doe                 |\n",
      "|---------------------|---------------------------|--------------------------|\n",
      "| **AI/ML Expertise** | **Advanced**              | **Intermediate**         |\n",
      "|                     | - Lead AI Research at DataMinds, published 3 papers on NLP architectures.  | - Developed ML models for predictive maintenance. |\n",
      "|                     | - Improved accuracy by 25% through a novel attention mechanism. | - Experience with MLOps pipeline reducing deployment time by 60%. |\n",
      "|                     | - Proficient in Deep Learning (PyTorch, TensorFlow), NLP (Transformers, BERT, GPT). | - Skilled in TensorFlow and PyTorch, but less emphasis on research. |\n",
      "| **Cloud Infrastructure** | **Basic**               | **Advanced**             |\n",
      "|                     | - No significant cloud infrastructure experience noted. | - Architected microservices infrastructure using Kubernetes, led cloud migration project. |\n",
      "|                     | - Experience primarily in AI/ML applications rather than cloud technologies. | - Strong AWS and DevOps skills with Kubernetes and Docker. |\n",
      "| **Leadership**      | **Intermediate**          | **Advanced**             |\n",
      "|                     | - Led a research team of 3 PhD candidates. | - Led a team of 5 developers and mentored junior developers. |\n",
      "|                     | - Experience in collaboration but less formal team management experience. | - Conducted code reviews and team management, showcasing strong leadership skills. |\n",
      "\n",
      "### Summary of Ratings\n",
      "\n",
      "- **Jane Smith**: \n",
      "  - **AI/ML Expertise**: Advanced\n",
      "  - **Cloud Infrastructure**: Basic\n",
      "  - **Leadership**: Intermediate\n",
      "\n",
      "- **John Doe**: \n",
      "  - **AI/ML Expertise**: Intermediate\n",
      "  - **Cloud Infrastructure**: Advanced\n",
      "  - **Leadership**: Advanced\n",
      "\n",
      "### Evidence\n",
      "- **Jane Smith**: \n",
      "  - \"Led research team of 3 PhD candidates\"【4:0†source】.\n",
      "  - \"Published 3 papers on NLP architectures\" and \"Developed novel attention mechanism improving accuracy by 25%\"【4:0†source】.\n",
      "\n",
      "- **John Doe**: \n",
      "  - \"Led team of 5 developers on cloud migration project\" and \"Implemented MLOps pipeline reducing deployment time by 60%\"【4:1†source】.\n",
      "  - \"Architected microservices infrastructure using Kubernetes\"【4:1†source】. \n",
      "\n",
      "This matrix indicates that while Jane Smith has a stronger background in AI/ML, John Doe excels in cloud infrastructure and leadership.\n",
      "\n",
      "Citations:\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【4:0†source】\"\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【4:0†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【4:1†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【4:1†source】\"\n",
      "\n",
      "Analyzing resumes...\n",
      "\n",
      "Here are the most recent and relevant projects each candidate has led, including team sizes and outcomes:\n",
      "\n",
      "### Jane Smith\n",
      "- **Position**: AI Research Lead at DataMinds (2020-Present)\n",
      "  - **Project**: Development of novel NLP architectures.\n",
      "  - **Team Size**: Led a research team of 3 PhD candidates.\n",
      "  - **Outcomes**:\n",
      "    - Published 3 papers on NLP architectures.\n",
      "    - Developed a novel attention mechanism that improved model accuracy by 25%【8:1†source】.\n",
      "\n",
      "- **Previous Role**: ML Engineer at AITech (2018-2020)\n",
      "  - **Project**: Implementation of computer vision models for autonomous systems.\n",
      "  - **Team Size**: Collaborated with cross-functional teams.\n",
      "  - **Outcomes**:\n",
      "    - Reduced model inference time by 35%【8:1†source】.\n",
      "\n",
      "### John Doe\n",
      "- **Position**: Lead Developer at TechCorp (2019-Present)\n",
      "  - **Project**: Cloud migration project.\n",
      "  - **Team Size**: Led a team of 5 developers.\n",
      "  - **Outcomes**:\n",
      "    - Successfully migrated services to the cloud.\n",
      "    - Implemented an MLOps pipeline that reduced deployment time by 60%【8:0†source】.\n",
      "\n",
      "- **Previous Role**: Senior Software Engineer at InnovSoft (2016-2019)\n",
      "  - **Project**: Development of machine learning models for predictive maintenance.\n",
      "  - **Team Size**: Worked within a team of engineers.\n",
      "  - **Outcomes**:\n",
      "    - Improved system performance by 40%【8:0†source】.\n",
      "\n",
      "### Summary\n",
      "- **Jane Smith** has demonstrated strong research capabilities with significant contributions to NLP, while also leading a small team effectively.\n",
      "- **John Doe** has shown leadership in a technical environment, focusing on cloud infrastructure and achieving substantial improvements in deployment efficiency. \n",
      "\n",
      "Both candidates have led impactful projects in their respective fields, showcasing their abilities to drive results and innovate.\n",
      "\n",
      "Citations:\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【8:1†source】\"\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【8:1†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【8:0†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【8:0†source】\"\n",
      "\n",
      "Analyzing resumes...\n",
      "\n",
      "### Comparison of Candidates' Experience with ML Systems Deployment\n",
      "\n",
      "#### Jane Smith\n",
      "- **Role**: AI Research Lead at DataMinds (2020-Present)\n",
      "  - **Experience**: \n",
      "    - Developed novel attention mechanisms that improved model accuracy by **25%**.\n",
      "    - Led a research team focused on advancing NLP technologies, demonstrating strong expertise in model refinement and deployment strategies.\n",
      "    - **Evidence of Successful Implementation**: While specific deployment metrics are not detailed, her leadership in research and published outcomes indicate a strong ability to implement complex ML systems effectively【12:1†source】.\n",
      "\n",
      "- **Previous Role**: ML Engineer at AITech (2018-2020)\n",
      "  - **Experience**:\n",
      "    - Implemented computer vision models for autonomous systems.\n",
      "    - Achieved a **35% reduction** in model inference time, showcasing successful deployment and optimization of ML models.\n",
      "    - Collaborated with cross-functional teams to ensure smooth integration of ML systems【12:1†source】.\n",
      "\n",
      "#### John Doe\n",
      "- **Role**: Lead Developer at TechCorp (2019-Present)\n",
      "  - **Experience**: \n",
      "    - Implemented an MLOps pipeline that reduced deployment time by **60%**, highlighting his expertise in streamlining ML systems deployment.\n",
      "    - Led a team of 5 developers on a cloud migration project, indicating significant responsibility in managing ML systems in a cloud environment.\n",
      "    - **Evidence of Successful Implementation**: His work in deploying ML models and optimizing processes has led to measurable efficiency improvements, making him proficient in practical application【12:0†source】.\n",
      "\n",
      "- **Previous Role**: Senior Software Engineer at InnovSoft (2016-2019)\n",
      "  - **Experience**: \n",
      "    - Developed machine learning models for predictive maintenance, contributing to a **40% improvement** in system performance.\n",
      "    - Architected microservices infrastructure using Kubernetes, facilitating reliable and scalable deployment of ML systems【12:0†source】.\n",
      "\n",
      "### Summary of Findings\n",
      "- **Jane Smith** has demonstrated a strong foundation in research and model optimization, particularly in NLP and computer vision, with quantifiable improvements in accuracy and inference times. However, her direct experience with deployment processes is more focused on research outcomes than operational metrics.\n",
      "  \n",
      "- **John Doe**, on the other hand, has substantial experience with operationalizing ML systems through MLOps and has achieved significant reductions in deployment times. His leadership in cloud migration projects and improvements in system performance position him as a strong candidate for roles focused on ML systems deployment.\n",
      "\n",
      "Both candidates showcase valuable skills, but John Doe appears to have a more hands-on approach to deploying ML systems effectively.\n",
      "\n",
      "Citations:\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【12:1†source】\"\n",
      "\n",
      "• From assistant-aKknmxW5c0aCWplnxG78jzzD:\n",
      "  \"【12:1†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【12:0†source】\"\n",
      "\n",
      "• From assistant-1fiEGSdxX13NCkya3mrcVmDr:\n",
      "  \"【12:0†source】\"\n",
      "\n",
      "Analyzing resumes...\n",
      "\n",
      "Here's a final recommendation table summarizing the top strengths, growth areas, and risk factors for both candidates:\n",
      "\n",
      "### Final Recommendation Table\n",
      "\n",
      "| Candidate      | Top 3 Strengths                                   | Growth Areas                               | Risk Factors                               |\n",
      "|----------------|---------------------------------------------------|-------------------------------------------|-------------------------------------------|\n",
      "| **Jane Smith** | 1. **Advanced AI/ML Expertise**: Strong background in NLP and deep learning with quantifiable improvements (e.g., 25% accuracy increase).  | 1. **Cloud Infrastructure Knowledge**: Limited experience in cloud deployment and management. | 1. **Limited Operational Experience**: Primarily focused on research; may require time to adapt to production environments. |\n",
      "|                | 2. **Research Leadership**: Effective in leading research teams and publishing impactful findings. | 2. **Practical Deployment Experience**: Needs more hands-on experience with end-to-end ML system deployments. | 2. **Team Management Skills**: Less formal management experience, which may affect team dynamics in larger projects. |\n",
      "|                | 3. **Innovative Problem-Solving**: Demonstrated ability to develop novel solutions and improve existing models significantly. | 3. **Broader ML Applications**: Could benefit from exposure to a wider range of ML applications beyond NLP. | 3. **Adaptability to Fast-Paced Environments**: Transitioning from research to industry may present challenges. |\n",
      "| **John Doe**   | 1. **Strong Cloud Infrastructure Skills**: Proven experience in architecting cloud solutions and managing ML deployments. | 1. **Depth in AI/ML Research**: Could enhance understanding of cutting-edge research in AI/ML. | 1. **Over-reliance on Technical Skills**: May need to develop soft skills for better cross-team collaboration. |\n",
      "|                | 2. **Effective Leadership**: Strong ability to manage teams, mentor junior developers, and lead projects successfully. | 2. **Continuous Learning in AI/ML**: Should stay updated with the latest research and advancements in AI/ML. | 2. **Risk of Burnout**: Leading multiple projects may lead to workload challenges. |\n",
      "|                | 3. **Proven Track Record in Deployment**: Achieved significant improvements in deployment efficiency (e.g., 60% reduction in deployment time). | 3. **Integration of Advanced AI Techniques**: Opportunities to incorporate more advanced AI methodologies into projects. | 3. **Potential Resistance to Change**: Experience with established processes may hinder adaptability to new methods. |\n",
      "\n",
      "### Summary\n",
      "- **Jane Smith** possesses strong research and AI/ML skills but needs to enhance her practical deployment experience and cloud infrastructure knowledge.\n",
      "- **John Doe** excels in cloud infrastructure and leadership while demonstrating a proven track record in deploying ML systems, but he could benefit from further research exposure and soft skills development.\n",
      "\n",
      "This table provides a clear overview to help make an informed decision regarding each candidate's fit for potential roles in the organization.\n",
      "Cleaning up resources...\n"
     ]
    }
   ],
   "source": [
    "await analyze_resumes()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}