Sample Tech Stack 1: For a Large-Scale NLP Application with Knowledge Graph Integration on GCP
Context Representation & Storage
- Knowledge Graph: Google Cloud Knowledge Graph
- Vector Embeddings:
- Vertex AI Feature Store
- Consider Compute Engine or Vertex AI Workbench for open-source libraries (FAISS, Annoy, ChromaDB).
- Explore Vertex AI Matching Engine for managed vector search.
- Document Storage: Cloud Firestore or Cloud Storage with Vertex AI Search and Conversation.
- Relational Database: Cloud SQL (MySQL, PostgreSQL, SQL Server).
Context Acquisition & Processing
- NLP Libraries: Python on Compute Engine, Cloud Functions, or Vertex AI Workbench using spaCy, NLTK, Hugging Face Transformers. Consider Vertex AI Natural Language API.
- API Integration: Python’s `requests` on GCP compute services or Cloud Functions. Consider Apigee API Management.
- Stream Processing: Cloud Dataflow, Cloud Pub/Sub.
Memory & Caching
- In-Memory Cache: Memorystore (Redis and Memcached).
Infrastructure
Programming Languages
- Python (primary for ML/NLP), Java, Go, etc.
Machine Learning Frameworks
- Vertex AI (TensorFlow, PyTorch, scikit-learn).
Sample Tech Stack 2: For a Robotics Application Focusing on Environmental Context on GCP
Context Representation & Storage
- Spatial Data: Cloud Spanner, Cloud Storage.
- Object Databases: Custom schemas within Cloud Firestore or Cloud Spanner.
- Sensor Data Storage: Cloud Bigtable, Cloud Time Series Insights.
Context Acquisition & Processing
- Sensor Data Ingestion: Consider IoT Platform (replacement for Cloud IoT Core), custom pipelines with Cloud Functions or Cloud Dataflow.
- Computer Vision: Run OpenCV, PyTorch Vision, TensorFlow Vision on Compute Engine or use Vertex AI Vision.
- Sensor Fusion: Custom logic on Compute Engine or GKE.
Memory & State Management
- Memorystore (Redis).
- Cloud Firestore or Cloud SQL for persistent state.
Infrastructure
- Google Cloud Platform (GCP)
- Compute Engine
- GKE
- Edge TPU (for on-device ML).
Programming Languages
- Python, C++, Go.
Machine Learning Frameworks
Sample Tech Stack 3: For a Dialogue System with Personalized Context on GCP
Context Representation & Storage
- User Profiles: Cloud Firestore or Cloud Spanner.
- Dialogue State Tracking: Application logic on Compute Engine, Cloud Functions, or Cloud Run with Memorystore (Redis) for caching.
- Conversation History: Cloud Firestore or Cloud Bigtable.
Context Acquisition & Processing
- NLP Libraries: Python on GCP compute with Rasa, spaCy, NLTK. Consider Vertex AI Natural Language API.
- STT & TTS: Cloud Speech-to-Text, Cloud Text-to-Speech.
- User Authentication: Firebase Authentication or Cloud IAM.
Memory & Caching
- Memorystore (Redis).
Infrastructure
Programming Languages
- Python, Node.js, Go.
Machine Learning Frameworks
