These skills do not affect the main grade but show the breadth of your profile.
AI Coding Assistants
▼
Uses ChatGPT/Claude for ML task assistance: debugging code, explaining algorithms, generating boilerplate. Formulates clear questions.
Applies ChatGPT/Claude for ML architecture discussions, code review, research. Uses for analyzing ML papers and documentation.
Defines policy for LLM usage in ML development. Evaluates risk/benefit. Establishes guidelines for the team.
Uses Cursor for ML code autocompletion: data pipeline generation, preprocessing functions, experiment boilerplate. Applies suggestions when working with pandas, numpy, scikit-learn. Verifies correctness of generated data transformations.
Effectively works with Cursor in ML projects: generates feature engineering pipelines, experiment configurations, test scenarios for models. Uses project context (notebooks, configs) for precise suggestions. Critically evaluates proposed model architectures.
Establishes AI tool usage policy in the ML team: what to generate (boilerplate, configs), what requires manual work (model architectures, loss functions). Evaluates AI assistant impact on experimentation speed.
Uses Copilot for generating ML code: boilerplate, data processing, feature engineering. Checks and corrects Copilot suggestions.
Effectively uses Copilot for ML development: complex feature engineering, sklearn pipelines, test generation. Writes context-rich comments for better suggestions.
Defines Copilot usage policy in ML team. Evaluates impact on productivity. Configures Copilot workspace for ML.
Alerting & On-Call
▼
Understands SLI, SLO and SLA concepts and their application to ML services. Knows typical SLIs for inference: latency p50/p95/p99, throughput, error rate and model prediction quality.
Defines SLIs and SLOs for ML services: inference latency, model freshness, prediction accuracy and availability. Configures SLI monitoring with alerts on approaching SLO violation for production models.
Defines SLI/SLO/SLA standards for organizational ML services considering downstream business impact. Designs reliability management architecture for ML platform with error budget policies and capacity planning.
API Management
▼
Reads and understands ML framework documentation: PyTorch, TensorFlow, scikit-learn and inference servers. Documents ML model APIs with input/output format descriptions and constraints.
Creates documentation for ML services: model API description, feature format, prediction response schema. Documents model cards with metrics, limitations and intended use for each model.
Defines ML ecosystem documentation strategy for different audiences: data scientists, engineers, product managers. Organizes documentation-as-code process with auto-generated model cards and API docs.
API Protocols
▼
Understands WebSocket protocol and its differences from HTTP for real-time communication. Knows WebSocket application for streaming inference and real-time ML task status updates.
Implements WebSocket API for streaming ML model inference results: token-by-token LLM generation, progressive image generation. Configures heartbeat, reconnection and message buffering for reliable delivery.
Defines WebSocket API architectural standards for ML platform: connection management, backpressure and failover. Designs scaling strategy for real-time ML services supporting thousands of concurrent connections.
Application Security
▼
Knows main OWASP Top 10 categories and their applicability to ML services and model APIs. Understands ML system threats: adversarial attacks, model poisoning and training data leakage.
Applies OWASP recommendations when developing ML model APIs: input validation, rate limiting, injection protection. Introduces security checks for model input data and training pipeline.
Defines security baseline for organizational ML systems based on OWASP ML Security and MITRE ATLAS framework. Designs ML pipeline protection architecture: secure training, model signing and inference monitoring.
Knows basic secure coding principles and their application in ML development. Does not hardcode API keys to model registries, validates inference request input data.
Implements secure data handling in ML pipelines: sensitive feature encryption, model access management. Applies data masking and anonymization for PII in training datasets.
Defines secure ML development standards for the organization: model governance, data lineage and access control. Designs ML platform security architecture with encryption, RBAC and compliance enforcement.
Authentication & Authorization
▼
Understands JWT/OAuth2 basics for ML services: API authentication for model endpoints, service-to-service token exchange for training pipelines, and OAuth2 scopes for data access control. Follows team practices for securing model serving APIs and experiment tracking services.
Implements OAuth 2.0 for ML platform: data scientist authentication, API keys for inference and service accounts. Configures scope-based authorization: access to models, data and experiments by roles.
Defines identity management standards for ML ecosystem: user, service and model authentication. Designs access architecture for sensitive data and models with audit trail and compliance.
Background Jobs & Task Queues
▼
Understands the fundamentals of Task Queues. Applies basic practices in daily work. Follows recommendations from the team and documentation.
Uses Celery for ML task orchestration: training, batch inference, feature computation. Configures priorities and retry policies. Monitors tasks through Flower.
Defines async ML processing strategy. Evaluates Celery vs Dramatiq vs Ray for ML workloads. Designs task orchestration for ML platform.
Batch Processing
▼
Effectively uses pandas for ML: data loading, EDA, feature engineering. Knows main operations: groupby, merge, pivot. Understands dtypes for memory optimization.
Optimizes pandas code for ML: vectorized operations, category dtype, chunked reading. Uses Polars for faster processing. Writes efficient feature engineering pipelines.
Defines data processing standards for ML team. Creates feature engineering framework. Trains the team on efficient data handling.
Defines data processing strategy for ML platform. Evaluates novel data processing frameworks. Designs unified data API for ML.
Caching
▼
Understands basics of Redis for caching ML feature data and model inference results. Reads cached feature vectors and prediction outputs for real-time serving optimization. Follows team guidelines for cache TTL settings aligned with model refresh cycles and data freshness requirements.
Uses Redis for caching ML predictions and features. Configures TTL strategies for ML cache. Applies Redis as online feature store for real-time serving.
Defines caching strategy for ML platform. Designs cache invalidation for model updates. Coordinates Redis infrastructure with DevOps.
CI/CD
▼
Understands basic GitHub Actions for ML: running model validation pipelines, reading CI logs for training test failures, understanding workflow configurations for ML experiment tracking. Follows team conventions for ML code CI automation.
Configures CI/CD for ML: tests, data validation, model evaluation. Uses matrix strategy for testing on different Python/CUDA versions. Caches ML dependencies.
Defines CI/CD standards for ML organization. Designs ML deployment pipeline. Coordinates CI/CD practices with DevOps.
Classical Machine Learning
▼
Trains XGBoost/LightGBM/CatBoost models with default parameters. Understands gradient boosting concept. Uses feature importance for model analysis.
Performs hyperparameter tuning for gradient boosting (learning_rate, max_depth, n_estimators, regularization). Handles categorical features (CatBoost native, target encoding). Configures early stopping and cross-validation. Analyzes SHAP values.
Defines gradient boosting usage strategy in ML organization. Evaluates gradient boosting vs deep learning for tabular data. Creates AutoML pipeline.
Defines tabular ML strategy for the organization. Researches novel gradient boosting approaches. Publishes results at conferences.
Clean Code & Refactoring
▼
Writes readable Python code for ML. Follows PEP 8 and uses black/ruff for formatting. Writes docstrings for functions. Understands the importance of reproducibility in ML.
Applies type hints in ML code. Uses mypy for static analysis. Writes unit tests for data processing and model evaluation. Organizes ML code into modules (data, features, models, evaluation).
Shapes code quality culture in ML team. Introduces code review practices for ML. Standardizes approaches to ML component testing.
Defines code quality strategy for ML organization. Creates ML-specific coding standards. Introduces industry best practices.
Cloud Providers
▼
Uses basic AWS services for ML: S3 for data, EC2 for training. Understands SageMaker basics. Works with AWS CLI for managing ML resources.
Uses AWS ML services: SageMaker for training, ECR for ML images, ECS/EKS for serving. Configures IAM for ML workloads. Uses S3 for model artifacts and data versioning.
Defines AWS strategy for ML platform. Designs multi-account ML architecture. Optimizes cloud costs for ML workloads. Coordinates with Cloud Engineering.
Defines cloud strategy for enterprise ML. Evaluates multi-cloud vs single-cloud for ML. Designs cloud-native ML platform architecture.
Concurrency & Parallelism
▼
Understands async/await basics in Python. Uses asyncio for asynchronous HTTP requests. Knows the difference between sync and async approaches for I/O-bound tasks.
Uses asyncio for asynchronous data preprocessing and API calls. Applies aiohttp for parallel requests to model serving endpoints. Writes async data pipelines.
Defines asynchronous programming standards for ML services. Designs async ML platform architecture. Evaluates async vs sync for different ML scenarios.
Data Modeling
▼
Understands basic data modeling principles: entities, relationships and normalization. Knows data structures for storing ML metadata: experiments, models, datasets and metrics.
Designs data schemas for ML platform: feature store schema, experiment metadata, model registry and prediction logs. Applies patterns for data and model versioning with lineage tracking.
Defines data modeling standards for organizational ML ecosystem: feature schemas, model signatures, dataset catalogs. Designs unified data model linking experiments, models and production metrics.
Database Optimization
▼
Understands how database indexes affect data loading speed for ML training pipelines. Knows basic indexing concepts for feature stores and experiment tracking databases. Follows team practices on index creation for frequently queried model metadata tables.
Creates composite indexes for optimizing feature extraction queries. Uses EXPLAIN for query plan analysis. Understands B-tree vs GIN indexes for different ML data types.
Defines indexing strategy for ML data. Coordinates with DBA on ML workload optimization. Monitors and optimizes unused indexes.
Writes efficient SQL queries for data extraction. Avoids SELECT *. Uses LIMIT for testing queries. Understands SQL execution order.
Analyzes slow queries through EXPLAIN ANALYZE. Optimizes feature engineering SQL: window functions, CTEs, subqueries. Uses batch operations for data loading.
Defines SQL optimization standards for ML team. Monitors and optimizes slow queries in ML pipelines. Creates guidelines for efficient SQL for data scientists.
Distributed Tracing
▼
Knows observability and distributed tracing concepts for ML services. Understands OpenTelemetry role in monitoring inference latency and throughput of ML models in production.
Instruments ML services with OpenTelemetry: traces for inference pipeline, metrics for model performance. Configures custom spans for tracking preprocessing, inference and postprocessing stages.
Defines ML observability standards based on OpenTelemetry with semantic conventions for ML context. Designs architecture for monitoring model drift, data quality and prediction confidence in production.
E2E Testing
▼
Understands end-to-end testing concept and its role in verifying ML pipelines from data to prediction. Writes simple E2E tests: data loading, preprocessing, inference and response format verification.
Creates E2E tests for ML pipeline: data ingestion, feature engineering, model training and serving. Verifies end-to-end flow from raw data to prediction API response with quality metric validation.
Defines E2E testing strategy for organizational ML systems covering training, serving and monitoring. Establishes standards for model release testing: performance benchmarks, fairness checks and safety evaluation.
GraphQL
▼
Understands basics of GraphQL schema design for ML experiment tracking and model registry APIs. Writes simple queries for model metadata, training metrics, and dataset information. Follows team conventions for exposing ML pipeline states through GraphQL interfaces.
Designs GraphQL schemas for ML platform: models, experiments, datasets, predictions with typing. Implements resolvers with integration to model registry, feature store and experiment tracker.
Defines GraphQL API architecture for organizational ML platform with federation for unifying data from different subsystems. Designs schema governance considering ML platform evolution and backward compatibility.
gRPC
▼
Understands basics of gRPC & Protocol Buffers for ML pipeline service communication. Reads proto definitions for training job orchestration and model serving schemas. Follows team conventions for defining feature store access contracts and experiment tracking service messages.
Creates gRPC services for ML inference. Defines protobuf schemas for ML predictions. Uses gRPC for internal communication between ML services. Configures streaming for batch inference.
Defines gRPC strategy for ML infrastructure. Designs service mesh for ML serving. Coordinates protobuf schema evolution.
Infrastructure as Code
▼
Understands Infrastructure as Code concept and basic Terraform syntax. Can read configurations for deploying ML infrastructure: GPU instances, S3 buckets for data.
Describes ML infrastructure in Terraform: SageMaker endpoints, EMR clusters, GPU EC2 instances and S3 data lakes. Creates modules for reusable ML environments with IAM and VPC configuration.
Defines IaC standards for organizational ML infrastructure with cost controls and compliance for GPU resources. Designs self-service platform for data scientists with automated ML environment provisioning.
Integration Testing
▼
Understands the need for integration tests in ML. Tests end-to-end pipeline on small dataset. Verifies that model can be loaded and return prediction.
Writes integration tests for ML pipeline: data → features → model → prediction. Tests API endpoints with real model. Uses test datasets for consistency.
Defines integration testing standards for ML. Designs test environments for ML. Coordinates testing between ML and backend teams.
Memory Management
▼
Understands basic memory management concepts: stack, heap, garbage collection and memory leaks. Knows OOM issues when working with large datasets and models in ML pipelines.
Optimizes memory usage in ML code: lazy data loading, batch processing, memory-mapped files. Manages GPU memory: gradient checkpointing, mixed precision and dynamic batch sizing.
Defines memory management standards for ML infrastructure: GPU memory budgets, OOM prevention and monitoring. Designs memory management architecture for distributed training with unified memory across nodes.
Message Queues & Event Streaming
▼
Understands basic Kafka concepts for ML pipelines: consuming training data streams, feature event ingestion, and basic producer/consumer patterns for model prediction events. Follows team practices for integrating ML services with Kafka-based data pipelines.
Uses Kafka for streaming feature computation. Configures producers/consumers for ML pipeline. Integrates Kafka with feature store for online features. Understands exactly-once semantics.
Defines Kafka usage strategy in ML infrastructure. Designs event schemas for ML data. Coordinates with data engineering on streaming pipelines.
Metrics & Monitoring
▼
Views dashboards in Grafana. Understands what metrics are (CPU, memory, latency). Can find the right dashboard and analyze a simple chart.
Adds custom metrics to application (counter, gauge, histogram). Writes PromQL queries for dashboards. Creates Grafana dashboards. Configures basic alerts (high error rate, high latency).
MLOps
▼
Understands ML pipeline concept: data → features → training → evaluation → deployment. Writes simple pipeline scripts. Uses Airflow DAG for basic ML workflow.
Designs ML pipelines with Kubeflow/Airflow. Configures parameterized pipelines for different models. Automates retraining with data quality checks. Implements pipeline testing.
Defines ML pipeline strategy. Standardizes pipeline components. Designs pipeline templating for faster development.
Defines ML pipeline platform strategy. Evaluates pipeline orchestrators. Designs enterprise ML pipeline architecture.
Uses MLflow for experiment logging: parameters, metrics, artifacts. Compares experiments in MLflow UI. Saves models through mlflow.log_model.
Designs MLflow workflow: experiment naming, run tags, artifact storage. Uses Model Registry for versioning. Configures autologging for sklearn/PyTorch. Writes custom MLflow Plugins.
Defines experiment tracking strategy for the organization. Evaluates MLflow vs W&B vs ClearML. Designs model governance workflow. Standardizes ML lifecycle.
Defines ML platform strategy. Designs enterprise experiment tracking. Evaluates and integrates ML lifecycle tools.
Deploys ML model as REST API through web framework/Flask. Understands inference pipeline: preprocessing → prediction → postprocessing. Uses pickle/joblib for model serialization.
Uses model serving frameworks: Triton, BentoML, Seldon. Configures batch and real-time inference. Optimizes inference latency (ONNX, model optimization). Configures A/B testing for models.
Defines model serving strategy for the platform. Designs unified serving layer. Optimizes serving costs. Coordinates with DevOps on infrastructure.
Defines enterprise model serving strategy. Evaluates serving technologies. Designs multi-model serving platform.
Networking
▼
Knows basic network protocols: TCP/IP, HTTP, DNS and their role in ML service operation. Understands network latency impact on inference latency and distributed training throughput.
Configures ML service networking: gRPC for inference, NCCL for distributed training, S3 for data. Optimizes network throughput for large dataset loading and gradient exchange.
Defines ML infrastructure network architecture considering bandwidth requirements for training and inference. Designs network topology for GPU clusters minimizing latency during all-reduce operations.
OOP & Design Patterns
▼
Knows basic patterns: Factory, Strategy, Observer. Understands how sklearn uses patterns (Pipeline, Transformer). Uses existing patterns in ML code.
Applies patterns in ML context: Strategy for model selection, Factory for feature pipeline creation, Observer for monitoring. Writes configurable ML components using Template Method.
Establishes ML system design patterns at team level. Defines reference architectures for common ML tasks. Trains the team on applying patterns in ML context.
Prompt Engineering
▼
Uses prompt engineering for model evaluation and dataset generation. Formulates prompts for code generation of training pipelines and data preprocessing. Understands prompt-based approaches to model testing and validation.
Applies advanced prompt techniques: chain-of-thought, few-shot, system prompts. Optimizes prompts for ML tasks: data analysis, code generation, debugging.
Defines prompt engineering practices for ML team. Introduces prompt-based automation in ML workflow. Standardizes approaches.
Defines prompt engineering strategy for the organization. Researches novel prompt techniques for ML. Evaluates impact on productivity.
Relational Databases
▼
Writes basic SQL queries for extracting data from PostgreSQL. Loads data into pandas via psycopg2/SQLAlchemy. Understands basic JOIN and GROUP BY.
Writes complex SQL queries for feature engineering. Uses window functions for time-series features. Optimizes queries with EXPLAIN. Creates materialized views for feature computation.
Defines PostgreSQL usage strategy in ML infrastructure. Designs data architecture for ML metadata. Coordinates with DBA on ML workload optimization.
System Design
▼
Understands basic system design principles and their application to ML infrastructure. Knows ML pipeline components: feature store, training pipeline, model registry and serving infrastructure.
Designs ML systems with separation into training and serving components considering scalability. Applies batch prediction, online inference and feature serving patterns for common ML tasks.
Defines system architecture of organizational ML platform, including feature platform, experiment tracking and model governance. Makes strategic decisions on trade-offs between latency, cost and model quality.
Type Systems
▼
Understands type safety concept and its importance for ML code reliability. Uses type hints in Python for annotating preprocessing functions and feature types in ML pipelines.
Applies strict typing in ML code: Pydantic models for API schemas, typed dataclasses for features. Creates type-safe interfaces for model inference with input/output schema validation.
Defines type safety standards for organizational ML codebase with mypy strict mode and pre-commit hooks. Designs type-safe SDK for ML platform with compile-time experiment configuration verification.