Garry Singh

Dynamic Principal AI and Data Engineer with 10+ years of hands-on experience architecting scalable big data pipelines, cloud-native infrastructures, and AI/ML-driven solutions for regulated industries including finance, government, retail, and compliance. Expert in real-time streaming, data governance, anomaly detection, and generative AI (LLMs, RAG, LangChain, LangGraph) to optimize performance, reduce latency by up to 60%, and ensure regulatory compliance. Proven leader in cross-functional teams, delivering high-impact projects with quantifiable results such as 30% error reduction and processing 10M+ daily metrics. 

Head of Product & Engineering

RegCore.AI – Regulatory Intelligence, Toronto, ON | December 2024 – Present

• Directed product and engineering roadmap for AI-powered regulatory platform, leveraging LLMs (OpenAI GPT, Anthropic Claude, Cohere), RAG, LangChain, and LangGraph for automated obligation mapping and explainable compliance workflows, slashing manual triage by >60% in pilots with financial institutions
• Architected cloud-native systems on Azure and AWS with Kubernetes, Docker, and Terraform for secure data flows, vendor integrations, and governance frameworks, incorporating knowledge graphs and agentic AI for proactive risk mitigation
• Led multidisciplinary teams (data scientists, engineers, legal experts) in PoC development, utilizing PySpark, TensorFlow, and Kafka Streams for real-time monitoring, achieving audit-ready traceability and stakeholder alignment
• Implemented MLOps pipelines with CI/CD (Jenkins, Azure DevOps) and monitoring (CloudWatch, ELK Stack) to iterate on client feedback, enhancing platform scalability and reducing deployment times

Principal Software Engineer

IBM / Employment and Social Development Canada (ESDC), Toronto, ON | August 2022 – April 2024

Project: Benefits Delivery Modernization and Old Age Security (BDM/OAS)

• Engineered ML models using Python, TensorFlow, PyTorch, and PySpark on Azure AKS for real-time anomaly detection in benefit transactions, processing millions of payments daily with 30% reduction in data outliers via YData-Profiling and data governance metrics
• Designed microservices with Java Spring Boot, Azure ADF, APIM (OAuth 2.0, TLS), and Confluent Kafka for guaranteed-delivery patterns, integrating Apache NiFi and Databricks for seamless data workflows
• Developed data integrity reports with Pandas, Matplotlib, and SQLAlchemy, extracting insights from healthcare/financial datasets while deploying ELK Stack for performance monitoring and issue resolution
• Orchestrated ETL/ELT processes with Prefect and Airflow, ensuring compliance with regulatory SLAs through lineage tracking and quality scoring in Snowflake and PostgreSQL (pgvector for vector embeddings)

Lead Senior Software Engineer

Northern Trust, Toronto, ON | April 2021 – August 2022

Project: GRM Enterprise Market Risk

• Built scalable data flows using Java, Scala, and Spark Structured Streaming on AWS EMR, processing 10M+ position metrics daily, optimizing risk reporting accuracy with Delta Lake on S3 and Parquet/ORC formats
• Optimized queries in SQL Server and Cassandra, reducing latency in high-frequency trading via AWS Glue and Lambda, and integrated SageMaker for predictive modeling of liquidity risks
• Developed REST APIs on AWS API Gateway with Swagger/OpenAPI, exposing enriched data for consumption, while implementing Bedrock and Guardrails AI for secure, compliant AI-driven calculations
• Engineered risk engines with Python (NumPy, Scikit-Learn) and Java (Spring Batch), leveraging CloudWatch for monitoring and Terraform for infrastructure-as-code, cutting processing times significantly

Senior Solutions Engineer

TD Bank, Toronto, ON | March 2020 – April 2021

Project: Convertible Trading Platform (CTP) - Capital Markets

• Migrated rule engines from Drools/Java to Scala (Akka, Play) with Druid, reducing infrastructure costs; implemented Spark SQL validations on Confluent Kafka topics for AWS-based in-memory processing
• Created Python ETL scripts (Pandas, NumPy) for data transformations (JSON, CSV, XML), integrating with AWS Data Lake and utilizing Kinesis for real-time synchronization of 5M+ transactions
• Orchestrated middleware migration from Solace to Confluent Kafka with connectors, KStreams, and KSQL, enhancing throughput and replay capabilities for audit compliance
• Built resilient Lambda microservices in Java 11, incorporating ReAct patterns and agentic AI for dynamic trading event handling

Senior Data Engineer / Lead Kafka Designer

Albertsons, Vancouver, BC | January 2019 – March 2020

Project: Enterprise Data Integration Services

• Programmed microservices in Java Spring with Confluent Kafka and Spark SQL for real-time validations, integrating StreamSets, KTables, and KSQL to MongoDB on Azure for retail data channels
• Engineered pipelines on Azure ADF for high-throughput integration from Blob Storage and Kafka, using xmltodict and fastavro in Python for schema-adherent transformations
• Deployed containerized architectures with Kubernetes, Docker, and Azure DevOps/ARM templates, embedding security (Kerberos, ACLs) and monitoring (Splunk) for governance
• Automated XML/Avro data orchestration, reducing errors through data profiling and quality checks

Software Engineer

Royal Bank of Canada, Toronto, ON | September 2014 – January 2019

Project: CTHUB - Corporate Treasury

• Developed Java/Struts applications with Spark for parallel processing of Basel III/OSFI reports, utilizing Hadoop (HDFS, MapReduce) and Hive for auditable data flows
• Designed ETL jobs in DataStage and Talend for ingesting diverse sources, optimizing storage on GCP with partitioned Parquet, BigQuery, and Cloud SQL
• Built secure APIs with Swagger on AWS API Gateway, integrating Oozie workflows and Flume/Sqoop for data-at-rest processing
• Implemented big data solutions with Storm and NiFi, enhancing traceability and compliance evidence retention

Master of Science in Software Engineering (AI & Machine Learning)

Oxford University, Oxford, UK

Dissertation: Real-Time Adaptive Knowledge Graph Construction for Regulatory Data – A Hybrid Approach Integrating Process Knowledge Learning/Forgetting Models and Large Language Models

Bachelor of Applied Science in Software Development & Network Engineering

Sheridan College, Toronto, ON

Bachelor of Business Administration

Vinayaka Missions Sikkim University, Sikkim, India