Anmol Dadlani — Data Engineer

Experience

Jun 2025 — present running

Data Engineer · RealityMine

Architected and deployed AWS Managed Airflow pipelines orchestrating AI agent–driven synthetic data collection across iOS and Android devices.
Owned the migration from legacy Azkaban orchestration to Airflow, modernising the team's scheduling and dependency management stack.
Implemented MLflow to standardise experiment tracking, model versioning, and reproducibility across ML workflows.
Designed and led internal AI agent workshops on agent-assisted data quality monitoring, cutting false positive alerts by 82%.
Re-architected ML training infrastructure on AWS SageMaker: 12× training throughput at 1.5× compute cost — 24 h → 2 h.

Jan 2023 — May 2025 success

Data Engineer · Codeweavers / Cox Automotive

Spearheaded the rebuild of live event tracking with Snowplow across automotive finance lender platforms (Black Horse, BNP Paribas), strengthening data collection and regulatory reporting for FCA-regulated lending journeys.
Implemented dbt across lender-specific data models, streamlining high-volume financial transaction processing to the accuracy required for risk, pricing, and compliance reporting.
Built custom dbt materialisations bringing Apache Iceberg tables into GCP, cutting query costs by 65% on the calculation datasets used by tier-1 lenders such as BNP Paribas and Lloyds Banking Group.
Built Metabase visualisations that enabled data-driven decision-making across the organisation.
Ran internal talks and workshops, training software developers and sharing project insights.

Jul 2020 — Jan 2023 success

Software Engineer · Codeweavers / Cox Automotive

Joined as a placement student and grew into a full-stack developer (Angular, C# .NET, PostgreSQL, AWS) shipping features into customer-facing quotation and loan origination platforms for UK car finance lenders.
Mentored new joiners in code quality, security, and auditability practices suited to regulated finance.

Projects

MSc thesis · 2023 82%

Social bias in large language models

Measured stereotypical bias in image-captioning datasets: alignment between demographic groups and concepts, analysis of Flickr30k captions and images, and profession-related target terms from StereoSet.

Python · Pandas · NLTK · scikit-learn · Hugging Face

Read the report ↓

MSc coursework · 2023 88%

Data mining

Decision trees over categorical attributes, k-means and agglomerative clustering of wholesale purchasing patterns, and NLP sentiment classification of coronavirus tweets.

Python · NumPy · Pandas · NLTK · scikit-learn

BSc final year · 2022 82%

Error detection on websites

ML system alerting developers to code changes likely to cause production failures — a Raspberry Pi web-scraping bot feeding GCP, a supervised screenshot-classification model, and a Flask API on Heroku scoring any submitted URL.

Python · TensorFlow · Selenium · Flask · GCP · Computer vision

Education

MSc Artificial Intelligence — King's College London

Pass with Merit. Agents & multi-agent systems, computer vision, AI reasoning & decision making, AI planning, data mining, deep learning.

2022 — 2023

BSc Computer Science & Artificial Intelligence — Loughborough University

First Class Honours, with Diploma in Professional Studies. Mathematics for CS, OOP in Java, databases, advanced networking, human–computer interaction.

2018 — 2022

A Levels — Sale Grammar School Sixth Form

Chemistry, Biology, Mathematics. SAF approved during tenure at Codeweavers.

2016 — 2018

Skills

Orchestration & ML platforms

Airflow
MLflow
SageMaker AI
dbt
TensorFlow
Machine learning
Jupyter

Cloud & infrastructure

AWS
GCP
EMR Serverless
Athena
BigQuery
Cloud Functions
Cloud Scheduler

Languages & data

Python
SQL
C#
PostgreSQL
Apache Iceberg
PowerBI