MLE Staff Or Principal Profile Ready

Each job type/level combination has its own resume (PDF) and knowledge base (JSON). This profile is specifically tailored for MLE Staff Or Principal positions.

Resume & Profile Complete

Resume Document Uploaded

Extracted Profile Information Extracted

Click to expand

Personal information extracted from your resume PDF. This data is used to auto-fill job application forms.

Profile Configuration (JSON Format)

{
  "address": {
    "city": "Durham",
    "state": "NC",
    "street": "2723 Western Bypass APT 107",
    "zip": "27705"
  },
  "education": [
    {
      "achievements": [
        "Won First Place in Duke AI Hackathon 2024 - Created Flashback (chief architect), a multimodal LLM-powered Q\u0026A system for video content via Pinecone \u0026 Google Cloud (Storage, Vertex AI Embedding, Gemini)",
        "Secured first place (solo) with record-breaking score in Modeling Process \u0026 Algorithms in-class Kaggle competition with comprehensive Machine Learning pipeline using LightGBM, Optuna, and ensemble models"
      ],
      "degree": "Master\u0027s in Artificial Intelligence",
      "gpa": "3.9/4.0",
      "institution_name": "Duke University",
      "location": "Durham, NC",
      "study_period": {
        "end_date": "Expected Dec 2025",
        "start_date": "Aug 2024"
      }
    }
  ],
  "job_preferences": {
    "willing_to_relocate": "Yes"
  },
  "leadership_and_publications": [
    "Certified as Google Developer Expert in Machine Learning (since 2018). Served as Google Summer of Code mentor for TensorFlow team (2022). Translated Machine Learning Pipelines on TensorFlow Extended (2021)",
    "Authored Conversational AI with Rasa (2021), Rasa in Action (2021), and spaCy in Action (2024)",
    "Created AgentSilex (PyPI), a minimalist agent framework with intelligent handoffs, context management, and LiteLLM integration; powers production systems requiring full control and transparency over agent behavior"
  ],
  "notable_achievements": {
    "academic_achievements": [
      "Won First Place in Duke AI Hackathon 2024 - Created \u0027Flashback\u0027 (chief architect)",
      "Secured first place (solo) with record-breaking score in Modeling Process \u0026 Algorithms in-class Kaggle competition"
    ],
    "professional_achievements": [
      "Built production AI systems serving 1M+ users across automotive and e-commerce",
      "Generated $15M+ business impact through AI implementations",
      "Improved user satisfaction 56%\u219288% for voice assistant transformation",
      "RAG system achieving 100K+ daily queries with 92% satisfaction",
      "Reduced \"no answer\" responses from 15% to 3% in RAG system",
      "Achieved 95% accuracy \u0026 0.8s latency for multi-turn query rewriting"
    ],
    "publications_and_contributions": [
      {
        "title": "Conversational AI with Rasa",
        "type": "Book",
        "url": "https://www.amazon.com/dp/1801077053",
        "year": "2021"
      },
      {
        "title": "Rasa in Action",
        "type": "Book",
        "url": "http://www.broadview.com.cn/book/5958",
        "year": "2021"
      },
      {
        "title": "spaCy in Action",
        "type": "Book",
        "url": "http://www.broadview.com.cn/article/420544",
        "year": "2024"
      },
      {
        "title": "Machine Learning Pipelines",
        "type": "Translation",
        "url": "https://www.ituring.com.cn/book/2815",
        "year": "2021"
      }
    ],
    "recognition_and_certifications": [
      "Google Developer Expert in Machine Learning (since 2018)",
      "Google Summer of Code mentor for TensorFlow team (2022)"
    ]
  },
  "personal_info": {
    "email": "u1mail2me@gmail.com",
    "ethnicity": "Asian",
    "first_name": "Xiaoquan",
    "gender": "Male",
    "is_disabled": "No",
    "is_us_citizen": "No",
    "is_veteran": "No",
    "last_name": "Kong",
    "links": {
      "github": "https://github.com/howl-anderson",
      "linkedin": "https://linkedin.com/in/xiaoquankong",
      "portfolio": "https://www.xiaoquankong.ai"
    },
    "location": "Durham, North Carolina, USA",
    "phone": "9193582215",
    "visa_status": "F-1 student visa with OPT eligibility after graduation (Expected Dec 2025)"
  },
  "professional_summary": "ML Engineer (10+ yrs, Baidu/Alibaba), Google Developer Expert. Built production AI systems serving 1M+ users; generated $15M+ impact. Duke MEng AI, creating graduate RL/RLHF course, 2x 1st AI winner",
  "projects": [
    {
      "key_features": [
        "Developed and deployed VS Code Extension with web dashboard using OpenAI Agents SDK",
        "Adopted by 3 Duke MS programs (MIDS, AIPI, DESIGNTK) with 200+ users and 10K+ monthly interactions",
        "Full-stack architecture with Next.js + FastAPI + VS Code integration"
      ],
      "project_description": "AI-powered VS Code extension and educational platform",
      "project_name": "Qubit VS Code Extension",
      "project_period": "Aug 2024 - Present",
      "project_status": "Production",
      "project_url": "https://marketplace.visualstudio.com/items?itemName=DukeCREATE.qubit",
      "web_dashboard_url": "https://qubit.xiaoquankong.ai/"
    },
    {
      "key_features": [
        "Won First Place in Duke AI Hackathon 2024 as chief architect",
        "Built with Pinecone \u0026 Google Cloud (Storage, Vertex AI Embedding, Gemini)",
        "Multimodal processing for video content Q\u0026A"
      ],
      "project_description": "Multimodal LLM-powered Q\u0026A system for video content",
      "project_name": "Flashback",
      "project_period": "Oct 2024",
      "project_status": "Hackathon Winner",
      "project_url": "https://devpost.com/software/flashback-o0jzlk"
    }
  ],
  "technical_skills": {
    "ai_ml": [
      "RAG",
      "LLM Tuning (LoRA/SFT)",
      "RL (PPO/DQN/RLHF)",
      "Prompt Engineering",
      "Multi-modal",
      "Agents",
      "MCP"
    ],
    "frameworks": [
      "PyTorch",
      "TensorFlow",
      "JAX",
      "HuggingFace",
      "LangGraph",
      "OpenAI Agent SDKs",
      "Google Agent SDKs"
    ],
    "languages_data": [
      "Python (Advanced)",
      "TypeScript",
      "SQL",
      "Spark/Hadoop",
      "Vector DBs",
      "Elasticsearch"
    ],
    "mlops_cloud": [
      "TensorFlow Serving",
      "K8s",
      "Docker",
      "CI/CD",
      "A/B Testing",
      "Observability",
      "GCP/AWS/Azure"
    ]
  },
  "visa_note": "F-1 STEM OPT eligible through 2028",
  "work_authorization": {
    "require_sponsorship": "Yes",
    "work_authorization_us": "Yes"
  },
  "work_experience": [
    {
      "company_name": "Duke University",
      "employment_period": {
        "end_date": "Present",
        "start_date": "Aug 2024"
      },
      "job_title": "Research Assistant \u0026 RL Course Developer",
      "key_achievements": [
        "Developed and deployed Qubit VS Code Extension and AI educational platform with web dashboard (OpenAI Agents SDK) and analytics; adopted by 3 Duke MS programs, 200+ users, 10K+ monthly interactions",
        "Architected full-stack system (Next.js + FastAPI + VS Code) with RBAC, invitation workflows, usage tracking, and RAG pipeline; delivered complete user journey from onboarding to AI-assisted coding with real-time analytics",
        "Developing graduate-level Reinforcement Learning course (Spring 2026 launch); building implementations for RLHF/LLM alignment (HuggingFace TRL), Embodied AI (MuJoCo robotics), Decision Transformers, World Models, and Multi-Agent Systems (PettingZoo); integrating 20+ algorithms spanning tabular RL to deep RL"
      ],
      "location": "Durham, NC",
      "rl_course_url": "https://rl-course.xiaoquankong.ai/"
    },
    {
      "company_division": "Intelligent Vehicle Systems",
      "company_name": "Baidu",
      "employment_period": {
        "end_date": "Aug 2024",
        "start_date": "Oct 2022"
      },
      "job_title": "Staff Applied AI Engineer",
      "key_achievements": [
        "Led architecture and core components of an end-to-end RAG system for owners\u0027 manual Q\u0026A (100K+/day, 92% satisfaction), with optimized chunking, BAAI/bge-m3, and Vector DB (Milvus)+Elasticsearch hybrid search",
        "Improved RAG retrieval accuracy by 14% via bge-reranker-large re-ranking and enhanced coverage by 30% through context-aware chunk expansion; reduced \"no answer\" responses from 15% to 3%",
        "Fine-tuned Qwen-1.5-14B with LoRA/SFT for multi-turn query rewriting; achieved 95% accuracy \u0026 p95 latency 0.8s using HuggingFace PEFT; validated via A/B tests and integrated into the production rewrite stage",
        "Automated LLM tuning with HPO (learning rate, batch size, LoRA rank) over synthetic datasets; compared prompting, SFT, RLHF (HF TRL); codified selection criteria and promoted best variants for the rewrite stage",
        "Built LLM/RAG MLOps: experiment tracking (W\u0026B), model registry, A/B tests, automated rollback, prompt versioning; added drift \u0026 business-metric dashboards with alerts and release checklists for safer rollouts"
      ],
      "location": "China"
    },
    {
      "company_division": "Automotive Manufacturer",
      "company_name": "Geely",
      "employment_period": {
        "end_date": "Oct 2022",
        "start_date": "Aug 2018"
      },
      "job_title": "Director of Engineering",
      "key_achievements": [
        "Led voice assistant migration from rule-based to deep learning NLU (BiLSTM+CRF, TextCNN); executed gray release to control risk; kept dialog flows backward-compatible; improved user satisfaction 56%\u219288%",
        "Built and scaled 11-person cross-functional team (AI + backend engineers); established ML training program that transformed backend engineers into ML practitioners deploying production models",
        "Redesigned dialog system into orchestrated architecture with domain isolation and clear interfaces; enabled independent domain releases and 3x faster iteration via modular design",
        "Formalized MLOps pipeline (data versioning, auto-deploy, monitoring) driving BERT NLU to 90% accuracy; optimized inference via TensorFlow Serving achieving 25% latency reduction, 3x throughput",
        "Deployed across Geely Holding brands (Volvo, Polestar, Lotus) and Benz; powering 1M+ vehicles worldwide"
      ],
      "location": "China"
    },
    {
      "company_division": "Ele.me Food Delivery",
      "company_name": "Alibaba",
      "employment_period": {
        "end_date": "Aug 2018",
        "start_date": "Sep 2015"
      },
      "job_title": "Senior Applied AI Engineer",
      "key_achievements": [
        "Led 2 engineers and 3 interns; built an intelligent customer service system (dual recall: Word2Vec + Elasticsearch); reduced agent workload by 28% and saved $15.3M/yr at 300K daily queries",
        "Launched an internal NLU platform adopted by 3 teams; cut model release time from 4 weeks to 7 days"
      ],
      "location": "China"
    }
  ]
}

Store personal information, work authorization, and preferences for MLE Staff Or Principal job applications in JSON format.

MLE Staff Or Principal Profile Ready

Resume & Profile Complete

Resume Document Uploaded

Drag and drop PDF file here

Extracted Profile Information Extracted

Drag and drop JSON file here