01
Discover the signal
Understand the data, dependencies, business context and failure modes before choosing a platform pattern.
Source contracts Dependency map Risk boundaries
CACHEORBIT / AI Data Engineer · Deloitte USI
DATASYSTEMS+ INTELLIGENCE
I work across cloud data engineering, lakehouse platforms, orchestration, reliability and applied AI—turning complex workflows into systems that are observable, maintainable and ready for production.
$ focus --current
01Core focusData + AI
02Cloud ecosystemMicrosoft Azure
03Engineering stanceProduction first
01 / EXPERTISEFour disciplines.
Four disciplines.
One dependable system.
Four interconnected capability areas—designed as one production system rather than isolated technical skills.
02 / SYSTEM JOURNEYFrom raw signal
From raw signal
to dependable system.
A scroll-led view of how I move from ambiguous inputs to production-ready data and AI systems.
01
Discover the signal
Understand the data, dependencies, business context and failure modes before choosing a platform pattern.
Source contracts Dependency map Risk boundaries
02
Engineer the movement
Build repeatable ingestion, transformation and orchestration paths with explicit recovery behaviour.
Metadata driven Idempotent runs Quality gates
03
Add useful intelligence
Apply AI where evidence can be retrieved, evaluated and reviewed—never as an ungrounded black box.
Grounded context Structured output Human review
04
Keep it dependable
Make systems observable, supportable and ready for production through measurable operational signals.
Run telemetry Actionable alerts Support playbooks
01Discover the signal
02Engineer the movement
03Add useful intelligence
04Keep it dependable
03 / ENGINEERING STORIESEngineering stories,
Engineering stories,
built as living systems.
Scroll vertically to travel horizontally through platform, applied-AI and product engineering work.
04 / ARCHITECTURE PATTERNSPatterns designed
Patterns designed
to survive production.
Reusable system patterns for lakehouse foundations, AI-assisted operations and resilient orchestration.
01
Architecture pattern
Enterprise lakehouse foundation
A governed medallion-style platform with metadata-driven ingestion, reusable transformation conventions, auditability and workload-aware optimisation.
Ingestion framework Quality gates Operational metadata Support dashboards
02
Applied AI pattern
AI-assisted pipeline operations
An assistant that interprets failures, retrieves run context and technical knowledge, proposes probable causes and creates a structured next-action brief.
Context retrieval Evidence-backed diagnosis Human approval Action brief
03
Reliability pattern
Resilient orchestration framework
A parameterised orchestration layer with idempotent execution, retry strategy, dependency controls, alert routing and supportable runbooks.
Recovery paths SLA signals Dependency controls Runbooks
05 / CAPABILITY CONSTELLATIONA stack that moves
A stack that moves
with the problem.
The platform is never the point. The right combination of data, cloud, AI and delivery capabilities is.
Core data engineering
Python · SQL · PySpark · Apache Spark · ETL / ELT · Data modelling · Apache Airflow · dbt
Azure & lakehouse
Azure Databricks · Azure Data Factory · ADLS Gen2 · Delta Lake · Unity Catalog · Azure Synapse · Microsoft Fabric · Key Vault patterns
Streaming & event systems
Apache Kafka · Azure Event Hubs · Spark Structured Streaming · Change data capture · Near-real-time serving
Warehousing & analytics
Snowflake · SAP HANA · Dimensional modelling · Semantic layers · Power BI · Query optimisation
AI & automation
Generative AI · Agent workflows · RAG patterns · LLM integration · Structured outputs · Evaluation design
Delivery & operations
CI/CD · Git · Azure DevOps · GitHub Actions · Docker · Observability · Performance tuning · Incident analysis
06 / INTERACTIVE LABDon't just read.
Don't just read.
Defend the signal.
A fast, touch-ready arcade challenge: let trusted data flow, quarantine anomalies and keep the intelligence core online.
LIVE DATA DEFENCE ARCADE
LIVE DATA DEFENCE
Signal Rush
SCORE00000COMBO×0WAVE1/5TIME40s
CORE INTEGRITY
100%
STREAM
01 · INGEST
02 · MODEL
03 · SERVE
04 · OBSERVE
AICORE
trusted packets flow through tap anomalies
Signal Rush ready.
07 / ENGINEERING PRINCIPLES
What I optimise for.
Make failure states visible before optimising the happy path.
Prefer reusable platform capabilities over one-off pipeline logic.
Treat quality, security and observability as design inputs—not add-ons.
Use AI where evidence can be retrieved, evaluated and reviewed by a human.
Communicate architecture so engineers and stakeholders can act on it.
Current roleAI Data EngineerDeloitte USI
FocusData + AI systemsProduction reliability
Identity contextIndependent engineering portfolioPersonal work and professional perspective
08 · Secure contact channel
Drop a note through the terminal.
Send a professional message, technical question, collaboration request or introduction. The production backend delivers it directly to Gourav's private mailbox.
Contact runtime initialized
Messages are validated server-side, rate limited and delivered through the configured email provider.
Terminal output cleared.