I build data pipelines, analytical warehouses, validation systems, and geospatial event-processing workflows with reproducible tests, explicit data contracts, and measurable reconciliation.

Independently designed geospatial data platform built on 10,000 real NASA FIRMS detections using Python, Kafka, Spark, Airflow, PostgreSQL/PostGIS, FastAPI, and AWS EMR Serverless.

Verified the complete local platform with 1M replay events and processed 5M replay events on AWS at 8,734.80 events/second—with zero rejects, zero duplicates, replay-neutral scientific aggregates, and independent checksum verification across 82 Silver/Gold files.

Python · PySpark · Kafka · Airflow · PostGIS · FastAPI · AWS EMR Serverless · CloudFormation

PostgreSQL dimensional warehouse with 400,000 verified sales facts, customer and product SCD Type 2 processing, surrogate-key resolution, validation and quarantine, repeat-safe loading, and 20 executed analytical SQL queries. Automated tests and hosted CI passed.

Facts resolve the dimension version current at load time; event-time historical attribution is not implemented.

Deterministic local 100,000-order batch pipeline with Pandas validation, rejected-record persistence, transactional SQLite snapshot replacement, rerun reconciliation, three analytical SQL queries, charting, and logging. Six tests passed; hosted CI uses a 1,000-row full-pipeline fixture.

Full-snapshot local batch processing—not incremental ingestion, CDC, streaming, cloud orchestration, or production deployment.

MIT-licensed Python tool that generates responsive desktop/mobile and light/dark GitHub profile heroes from YAML. It produces animated GIFs, static reduced-motion fallbacks, ready-to-copy profile markup, and publishing-safety validation. Version 1.0.0 has 14 automated tests and passing hosted CI on Python 3.11 and 3.12.

Python · Pillow · PyYAML · GitHub Actions · Accessibility

• Languages: Python, SQL, Java, Scala, Shell
• Data platforms: Apache Spark, Apache Kafka, Airflow, Pandas, Hadoop/HDFS/YARN, Hive, Presto, AWS Glue, Informatica
• Streaming and CDC: Kafka Connect, Debezium, Amazon Kinesis, Amazon Data Firehose
• Cloud: AWS EMR Serverless, S3, Glue Data Catalog, Athena, Lambda, Redshift, IAM, KMS, CloudFormation, CloudWatch
• Databases and modelling: PostgreSQL, PostGIS, SQLite, dimensional modelling, star and snowflake schemas, SCD Type 2
• Reliability and delivery: validation, reconciliation, idempotency, retries and recovery, quarantine, Git, Jenkins, GitHub Actions, Docker, ELK
• Serving and visualization: FastAPI, Streamlit
• Web delivery: JavaScript, HTML/CSS, Node.js, Vercel, Supabase, Zod, responsive accessibility, security headers

Completed five NASA Applied Remote Sensing Training (ARSET) certificates of completion covering remote-sensing fundamentals, hyperspectral data for land and coastal systems, and sustainable Earth-science application development.

• Fundamentals of Remote Sensing — August 3, 2026
• Hyperspectral Data for Land and Coastal Systems — August 17, 2026
• Developing Sustainable Earth Science Applications, Modules 1–3 — August 19, 2026

• Evidence before claims and measurable validation before scale statements
• Reproducible execution with explicit data contracts and rerun behavior
• Reconciliation and failure handling treated as part of the pipeline
• Honest limitations and tradeoffs documented beside the implementation

• ASTRAYAN.dev
• LinkedIn
• nitheeshc2325@gmail.com
• Open to Data Engineering opportunities