Download 100 Projects to build Forward Deployed Engineers Portfolio Course for Free

CDC pipeline from PostgreSQL to BigQuery with late arrivals

Kafka → Delta Lake streaming with watermarking and compaction

Airflow Dagster pipeline deduping and enriching CRM leads

Data quality framework (Great Expectations) with auto-fail

FluentBit → S3 → Athena log pipeline with partition pruning

Ingest messy CSVs to Parquet lake with schema validation

Feature store (Feast Tecton) with point-in-time correctness

ELT reconciling Stripe and Shopify revenue nightly

Real-time IoT anomaly detection with Flink Spark

Data contract enforcement with JSON Schema + Pydantic

REST API gateway normalizing HubSpot, Salesforce, Notion

GraphQL facade over heterogeneous internal microservices

API versioning with backward-compatible deprecation

OAuth2 proxy unifying Google, Microsoft, GitHub SSO

Webhook relay with retry, DLQs, and idempotency keys

Rate-limiting and quota middleware for internal APIs

Bi-directional Airtable ↔ Postgres sync with conflicts

Connector for niche SaaS into data warehouse

Event-driven integration using NATS JetStream

Public API for internal tool with OpenAPI spec

MCP server for vector DB with typed tool schemas

MCP server exposing Jira Confluence to LLM tools

MCP server for Slack, Gmail, Calendar assistant agent

MCP server for Zendesk with auto-triage tools

MCP server wrapping internal pricing engine, audited

MCP server for BI tool enabling NL queries

MCP server for secrets manager with least-privilege

MCP server for document store with access control

MCP server for feature flags with rollout tools

MCP server for monitoring stack with incident tools

Basic RAG over internal docs with retrieval metrics

Multi-tenant RAG with isolation and access control

RAG with hybrid search and query rewriting

Agentic RAG with tool use and plan-and-execute

RAG with citations and grounded answers

RAG eval harness (recall, precision, faithfulness)

Retrieval with re-ranking for quality

LLM data doc generator from schemas and queries

Fine-tune small LLM for domain Q&A (e.g., SOX)

LLM agent drafting SQL with schema grounding

End-to-end RAG eval with golden datasets

Drift detection for embeddings and Recall@K

LLM response quality scorer (rules + model)

Cost attribution dashboard for LLM calls

Latency SLO dashboards for LLM endpoints

Canary eval suite for prompt changes (A B)

Offline online metric correlation for LLM changes

Red-team harness for jailbreaks and injection

Hallucination detector using claim verification

Feedback loop UI to flag bad answers, retrain