# K-M Samiul Haque

> Staff Software Engineer · AI/ML Systems · Platform Engineering

Staff Software Engineer at RBC with 9 years of continuous IC progression. Building LLM ensemble labelling pipelines, knowledge distillation systems, and sub-50ms intent classification at Canada's largest bank.

## Links

- Website: https://www.samiulhaque.com
- LinkedIn: https://www.linkedin.com/in/samiul-haque
- GitHub: https://github.com/samhaque
- Email: sammy.haque@alumni.utoronto.ca
- Resume: https://www.samiulhaque.com/static/content/K-M_Samiul_Haque_Resume_2025.pdf

## About

Nine years at Canada's largest bank. From intern to Staff. Always building.

Nearly a decade ago, I joined RBC as an intern with a Statistics degree and a curiosity about what production software really looks like at scale. That curiosity led me through six distinct engineering disciplines — ChatOps, SRE, security platforms, developer experience — before arriving at what I find most compelling: building AI systems that work at the intersection of research and production.

Today, I lead AI engineering for an end-to-end call center platform at Canada's largest bank. My current work spans LLM ensemble labelling pipelines, knowledge distillation of large embedding models, and intent classification systems routing 14,000+ banking intents with under 50 ms latency on CPU-only infrastructure. The challenge is bridging the gap between what's possible in an AI research lab and what can actually be deployed reliably at enterprise scale.

My foundation is a Statistics degree from the University of Toronto — which means I think in distributions, not just decisions. I care about why models behave the way they do, not just whether they hit a benchmark. That mathematical rigour shapes how I approach everything from embedding fine-tuning to production observability.

Outside work, I maintain a homelab, experiment with network engineering, and prototype ideas. I'm interested in the long game of technology: systems that age gracefully, teams that grow sustainably, and engineering that compounds.

## Current Role

**Staff Software Engineer** — AI Advice Centre, Royal Bank of Canada (Sep 2025 – Present)

- Architect end-to-end AI call center platform integrating SIP telephony, real-time speech-to-text transcription, and LLM summarization pipelines processing thousands of daily banking interactions.
- Engineered LLM ensemble labelling pipeline using GPT-4o mini, GPT-4.1 mini, and GPT-5 mini with consensus voting and tie-breaking escalation to larger thinking models; generated high-quality training labels at scale for fine-tuning.
- Implemented knowledge distillation workflow: used LLM ensemble as teacher to fine-tune Arctic Embed 2.0 Large as domain-specific student model for banking intent semantics.
- Built continuous training MLOps pipeline on S3 + Apache Airflow; automated dataset ingestion, model retraining, evaluation, and deployment to OpenShift.
- Benchmarked SVM (RBF kernel), logistic regression, SetFit (head and full fine-tune), centroid, and fine-tuned Arctic Embed classifiers for production intent routing.
- Led BART intent classification system using Jina v3 / multilingual-e5 with ONNX Runtime; reduced inference latency from 8 s → 50 ms on CPU-only OpenShift 4.
- Built Spring Boot microservices with k-NN routing of 14,000+ banking intents across 110 workflow categories; achieved 95% semantic naming accuracy via LLM-assisted clustering.

Tech stack: Python · Java · Spring Boot · ONNX Runtime · Arctic Embed 2.0 · Jina v3 · multilingual-e5 · SetFit · SVM · GPT-4/4o/5 · Claude · S3 · Apache Airflow · OpenShift 4 · AWS Transcribe/Bedrock

## Career History

- **Staff Software Engineer** — Security Platforms, Royal Bank of Canada (Jul 2024 – Sep 2025)
- **Lead Software Engineer** — SRE & Chaos Engineering, Royal Bank of Canada (Aug 2022 – Jul 2024)
- **Senior Software Developer** — Developer Experience & OSPO, Royal Bank of Canada (Nov 2020 – Aug 2022)
- **Software Developer** — Developer Experience, Royal Bank of Canada (Aug 2019 – Nov 2020)
- **Software Developer Intern → SRE Intern** — 3 terms, Royal Bank of Canada (May 2017 – Aug 2019)
- **Full-Stack Developer** — Contract, University of Toronto (Sep 2017 – Feb 2018)

## Key Projects

- **BART Intent Router**: LLM ensemble labelling pipeline (GPT-4o mini + GPT-4.1 mini + GPT-5 mini with consensus voting) used as teacher to fine-tune Arctic Embed 2.0 Large as a domain-specific student model. Routes 14,000+ banking intents across 110 categories with sub-50ms latency on CPU-only OpenShift 4. | Metrics: <50ms CPU inference, 14K+ intents, 95% naming accuracy — Internal/proprietary
- **AI Call Center Platform**: End-to-end AI-powered platform integrating SIP telephony, real-time speech-to-text transcription, and LLM summarization pipelines. Processes thousands of daily banking interactions with automated intent classification and agent-assist workflows. | Metrics: 1000s of daily interactions, Real-time SIP, Multi-model LLM — Internal/proprietary
- **Vulnerability Triage Platform**: Centralized vulnerability management platform spanning Canada, US, and UK subsidiaries. Unified data ingestion from Snyk, NexusIQ, Kenna, Tenable, and Aqua into a Snowflake warehouse with real-time triage dashboards. | Metrics: 60% faster assessments, 3 global subsidiaries, 5 security feeds — Internal/proprietary
- **Chaos Engineering Platform**: Automated large-scale chaos experiments across AKS, OpenShift, and VMware clusters using Tanium and Gremlin APIs. Enables self-service chaos scheduling for VM and Kubernetes workloads across multi-cloud infrastructure. | Metrics: 90% setup reduction, AKS + OpenShift + VMware, 100s of VMs/pods — Internal/proprietary
- **MirrML**: Flask + Clarifai Image Recognition API; clothing style classifier (business/casual/evening) via neural network trained on scraped image data. Matches users with friends with similar style profiles. — [https://github.com/samhaque/UofTHacksIV_MirrML](https://github.com/samhaque/UofTHacksIV_MirrML)
- **LendR**: Android app using NFC technology for social micro-financing. Built a Karma system to incentivize repayment; negative karma reduces borrowing limits. Custom backend tracks transactions and karma ratings. — [https://github.com/samhaque/FinTech_LendR](https://github.com/samhaque/FinTech_LendR)
- **HackTheValley API**: Production event management API for HackTheValley, a University of Toronto hackathon. Built with Go for high performance and deployed for the annual event. — [https://github.com/hackthevalley/htv-api](https://github.com/hackthevalley/htv-api)

## Skills

- **AI / ML Systems**: LLM Ensemble Labelling, Knowledge Distillation, Arctic Embed 2.0, Embedding Fine-tuning, ONNX Runtime, SetFit, SVM (RBF), Intent Classification, Semantic Clustering, RAG, Prompt Engineering, RASA NLU, Clarifai
- **Languages**: Python, Java, Go (Golang), JavaScript, TypeScript, SQL, Shell/Bash, C, C++, Ruby
- **MLOps & Data Pipelines**: Apache Airflow, S3 Data Pipelines, Continuous Training, Model Evaluation, Snowflake, Apache Kafka
- **Cloud & Containers**: OpenShift (OCP), Kubernetes, Docker, Azure AKS, AWS (Transcribe, Bedrock, S3), PCF
- **Frameworks**: Spring Boot, FastAPI, Flask, Django, React, Gin, ASP.NET Core, RabbitMQ
- **Databases**: PostgreSQL, SQL Server, Elasticsearch, MongoDB, Redis, MariaDB, MySQL, Oracle DB
- **DevOps & CI/CD**: GitHub Actions, Jenkins, Ansible, HashiCorp Vault, Artifactory, SonarQube, IBM UDeploy, Git
- **SRE & Observability**: Gremlin, Tanium, PagerDuty, ServiceNow, ELK Stack, Prometheus, Grafana, Moogsoft, Dynatrace
- **Security**: Snyk, NexusIQ, Kenna, Tenable, Aqua, Recorded Future, mTLS, PingFederate
- **AI Dev Tools**: Claude, ChatGPT / GPT-4o, GitHub Copilot, Cursor

## Education

- **Honours B.Sc., Statistics** — University of Toronto (2015 – 2021). Double Minor: Geographic Information Systems (GIS) & Psychology
- **Engineering Studies** — Ontario Tech University (2024 – Present)

## Volunteering

- **VP of Development** — AMACSS — Association of Mathematical & Computer Science Students (Jun 2018 – Apr 2020): Led development projects and technical initiatives for the mathematics and computer science student association at University of Toronto Scarborough.
- **Developer / SRE** — Computer Science Enrichment Club (CSEC, UofT) (Aug 2016 – Apr 2020): Contributed to club infrastructure and development projects, mentored students in software engineering and site reliability engineering fundamentals.
- **Web Developer** — Women in Computer Science, Statistics & Mathematics (WiCSM) (Aug 2020 – Present): Built and maintained the organization's web presence, supporting an initiative that advocates for gender diversity in STEM at the University of Toronto.

## Full Version

For complete details: https://www.samiulhaque.com/llms-full.txt