DATA
ENGINEER

AI, Data and Cloud enthusiast building large-scale data pipelines and agentic AI systems. Years of data engineering experience in Marketing & Business and Satellite data.

4+
Years in data
Geospatial
Satellite data
Business
Analytics data
4.0
MSc CS GPA
Top 1%
Undergrad rank

About Me

// who I am
Shrey Niraula

I am Shrey Niraula, a data engineer and computer science graduate. Growing up in Japan, I spent countless hours around games and computers, which sparked a curiosity about how software works and eventually inspired me to start building things of my own.

That curiosity guided me through Electronics Engineering and into a career centered on data, software, and AI. Today, I work at the intersection of data engineering and agentic systems, exploring how intelligent applications can combine reliable data foundations with autonomous reasoning and decision-making.

Having spent part of my childhood in Japan, I still enjoy keeping up with the language and culture. Outside of work, you will usually find me watching films and anime, listening to music, or diving into whatever new curiosity has captured my attention.

Nepali (native) Japanese (fluent) English (fluent)

Skills

// my stack
Programming Languages
Python
Scala
SQL
Cloud & Big Data
AWS (EC2, S3, Lambda)
Athena · Glue · DynamoDB · EMR · MSK
Spark
Apache Airflow
Databases & Caching
MySQL
Redshift
Dremio
Iceberg
Postgres
Redis
Software Engineering
Flask
Docker
RabbitMQ
Git
Terraform
RESTful APIs
React
Linux
CI/CD
OOP
Multithreading
Distributed Systems
Data Concepts
Data Lake
Data Warehousing
Data Modeling
SCD
Data Orchestration
Data Governance
ETL / ELT
OLAP
AI / ML
Ollama
LangGraph
RAG
Prompt Engineering
Vector Databases
Data Science
Pandas
NumPy
Matplotlib
Statistical Modeling

Experience

// where I have worked
Aug 2024 to Present Current

Data Engineer

Huntsville, AL, USA
NASA ODSI (CSDA Program)
  • Built and evaluated a multi-agent system on a local LLM that converts Amazon S3 inventory reports into auditable, cost-saving storage-migration plans, using verifier-driven retries for numerical accuracy; validated on petabyte-scale CSDA production buckets against a status-quo baseline.
  • Onboarded a new satellite vendor (Airbus DEM) end-to-end, building its ingestion and PySTAC-based STAC transformation from scratch.
  • Rebuilt legacy Airbus SAR and BlackSky ingestion into a centralized, partitioned Airflow pipeline, with Pytest suites validating ingestion logic.
  • Built an Airflow solution to automate the manual MAXAR vendor checksum-verification process, consolidating DynamoDB export and Athena checks into a single DAG.
Apache AirflowAWS (Athena, DynamoDB, S3, Glue)PySTACMetadata mappingPytest
May 2022 to Jul 2024

Mid-Level Data Engineer

Kathmandu, Nepal · Hybrid (US-based stakeholders)
GrowByData Services
  • Owned the LeadGen data source end-to-end (POC, acquisition, ETL, and reporting backend).
  • Re-architected the Facebook Ads pipeline end-to-end (acquisition, ETL, reporting) into a modular, multithreaded design supporting easy extension of pulls and filters.
  • Migrated the reporting warehouse from Redshift (ra3.4xlarge) to a Dremio lakehouse (S3 + Glue Catalog), decoupling storage and compute to scale compute independently for heavy ad-hoc/aggregation queries.
  • Refactored demo-report anonymization from existing duplicated query classes into a reusable Python decorator with pandas post-processing, cutting 100+ report development from ~3 months to ~2 weeks.
  • Sped up acquisition 2–5× across Pinterest, TikTok, Google/Bing Ads, and Shopify via multithreading and memory tuning.
  • Integrated Redis caching into the reporting layer, cutting query latency ~70%.
  • Mentored junior engineers and interns across data acquisition, ETL, reporting logic, Spark, and data modeling.
LakehouseDatalakeEMRGlueCrawlerRedisOptimizationETL OrchestrationDremio
May 2021 to May 2022

Associate Data Engineer

Kathmandu, Nepal · Hybrid (US-based stakeholders)
GrowByData Services
  • Built the first production acquisition pipeline for Pinterest Ads with OAuth2 authentication.
  • Refactored the Facebook Ads acquisition to handle API throttle limits, stabilizing ingestion.
  • Migrated Talend ETLs to distributed Spark (Scala + SparkSQL) using star-schema modeling and SCD, reducing Redshift CPU load and ETL latency.
Data WarehousingData ModelingTalendPySparkScalaSparkSQLRedshiftStar-Schema DesignOAuth2

Education

// where I studied
GPA 4.0

The University of Alabama in Huntsville

Aug 2024 to Aug 2026 Huntsville, AL, USA
Master's in Computer Science, Data Science Concentration

Thesis: Agentic Cloud Decision Framework, multi-agent local LLM orchestration and tool-augmented reasoning for large-scale NASA cloud storage.

Advanced DatabaseData Structures & AlgorithmsBig Data ComputingCloud ComputingMachine Learning
81.70% (Rank 4th, top 1%)

Tribhuvan University, IOE, Pulchowk Campus

Nov 2016 to Mar 2021 Lalitpur, Nepal
Bachelor's in Electronics and Communication Engineering

Ncell Excellence Cash Award (2019) for topping the consecutive 3rd and 4th semesters.

Object Oriented ProgrammingDatabaseArtificial IntelligenceData MiningLinear AlgebraComputer Architecture

Projects

// selected work
Self-Hosted Agentic LLM Framework
01 / 05
MS Thesis 2026

Self-Hosted Agentic LLM Framework

Multi-agentic systems via local LLMs, with typed outputs, deterministic verification, and a self-correction loop, producing auditable results.

Pydantic AIOllamaMulti-AgentText-to-SQLPython
Visual SLAM in Dynamic Scenes
02 / 05
Final Year 2020

Visual SLAM in Dynamic Scenes

Visual SLAM project to localize and navigate a mobile robot in a dynamic environment.

Visual SLAMOpenVSLAMROSICNetPyTorchSemantic Segmentation
Precision Livestock Farming
03 / 05
Locus Winner 2019

Precision Livestock Farming

A data-collection and analysis system that monitors broiler chickens to estimate welfare and automate their habitat.

YOLOSORTSound AnalysisIoTMobile AppAutomation
Traffic Management & Analysis
04 / 05
2019

Traffic Management & Analysis

An intelligent traffic system for Kathmandu Valley that detects flow, analyzes data, and optimizes signal timing across major intersections.

PythonFlaskJavaScriptAJAX / jQueryPTV VissimData CollectionSimulationData Modeling
05 / 05
2019

ABU Robocon 2019 Simulator

A graphics project that simulates ABU Robocon 2019 robots and the arena to find the best robot configuration for the competition.

C++OpenGLCamera CalibrationLighting EffectsModel Rendering

Participations

// hackathons & programs
International Hackathon 2020

Quantum Hack

First International Digital Hackathon in Nepal
AI Research 2019

Nepal Winter School in AI

AI-NAAMII Winter School, Pokhara
Hardware Systems 2019

Hardware Project Competition

Locus Nepal
Hackathon 2019

Everest Hack

Everest Hack Nepal
Assistive Tech 2018

Disaster Hack

AT-Hackathon Nepal
Innovation 2018

Locus 2018

Locus Nepal

Certifications

// courses I completed
AWS In Progress

AWS Certified Data Engineer - Associate (DEA-C01)

JEES Dec 2023

Japanese-Language Proficiency Test (JLPT) N2

Verify
Coursera Aug 2020

Front-End Web Development with React

Verify
Coursera Aug 2020

Server-side Development with NodeJS, Express and MongoDB

Verify
Coursera Jun 2020

SQL for Data Science

Verify
Coursera Jan 2020

Machine Learning (Andrew Ng)

Verify
edX Sep 2020

Applied Deep Learning Capstone Project (DL0320EN)

Verify

Contacts

// open to work and good conversations