Available for Roles · AI / ML · Agentic Systems · Computer Vision

Faisal
Sayed

AI Engineer · Generative AI Specialist · MSc Computing Science (AI)

~/faisal # specializations
$ echo
▶ 2+ yrs · Production AI Systems · LangChain · RAG · AWS · YOLOv8
92%
ASL Classifier Accuracy
35%
Query Resolution Speedup
18%
Detection Precision Gain
5+
Open-Source AI Projects
View Projects LinkedIn GitHub
// about

Building AI that
ships to production

I'm an AI/ML Engineer with 2+ years of hands-on experience building production-grade Generative AI systems, NLP pipelines, and computer vision solutions. Currently completing my MSc in Computing Science (AI Specialization) at the University of Alberta, affiliated with the Alberta Machine Intelligence Institute (Amii).

My work spans the full AI engineering lifecycle — from research ideation and model fine-tuning to scalable cloud deployment on AWS and IBM Cloud. I specialize in LLM-based agentic systems, RAG architectures, and real-time computer vision, with a track record of measurable business impact: 35% faster query resolution, 18% detection precision improvements, and 92% classification accuracy on challenging real-world datasets.

I'm fully authorized to work in Canada and open to relocation for the right opportunity — including FAANG-tier roles where AI systems are built at scale and for real-world impact.

// tech stack
Generative AI & LLMs
LangChainRAGGPTLLaMAMistralPrompt EngineeringFine-Tuning
ML Frameworks
PyTorchTensorFlowScikit-LearnKerasXGBoostHuggingFace
Computer Vision
YOLOv8OpenCVMediaPipeCLAHE
Cloud & MLOps
AWS SageMakerLambdaStep FunctionsMLflowDockerIBM Cloud
Languages
PythonSQLRMATLAB
// experience

Professional Journey

From data pipelines to production LLM systems — a progression across industry and research.

Independent AI Research & Freelance Jun 2025 – Present
AI Engineer · Remote
  • Built a real-time Sign Language to Text converter achieving ~92% character recognition accuracy across a 26-class ASL dataset using a custom NLP + computer vision pipeline.
  • Developed and published an open-source RAG-based document Q&A tool (LangChain + FAISS) supporting semantic search over 100+ page PDFs with sub-2s response times.
  • Completed AWS Certified Machine Learning – Specialty preparation coursework and hands-on labs.
  • Actively contributing to Kaggle competitions and open-source ML repositories; maintains 5+ documented AI/ML projects on GitHub.
University of Alberta Sep 2024 – Dec 2024
Research Assistant · Edmonton, AB
  • Designed an image enhancement pipeline (CLAHE + adaptive illumination) improving YOLOv8 object detection precision by ~18% on a low-light cattle image dataset.
  • Co-authored a research paper on autonomous image pre-processing for challenging lighting environments, submitted to a peer-reviewed venue.
  • Automated data collection, preprocessing, and model benchmarking across 3 dataset configurations, cutting evaluation time by 30%.
Accountium May 2024 – Aug 2024
Generative AI Developer · Vancouver, BC (Remote)
  • Deployed GPT and LLaMA-based conversational chatbots integrated into client-facing business systems, reducing average customer query resolution time by ~35%.
  • Applied prompt engineering and fine-tuning techniques, achieving a 20% improvement in user satisfaction scores in A/B testing.
  • Integrated chatbots with backend APIs and CRM platforms for 3 enterprise clients, ensuring compliance with PIPEDA.
  • Established automated QA pipelines for model drift monitoring, reducing manual review overhead by 40%.
University of Alberta Jan 2024 – Apr 2024
Teaching Assistant – Software Quality · Edmonton, AB
  • Delivered lab sessions and graded assignments for 80+ final-year Computing Science students on software quality and testing principles.
  • Provided one-on-one mentorship on debugging, code review, and TDD — contributing to a 15% improvement in average lab scores.
Bosch Jun 2022 – Jul 2023
Data Engineer · Mumbai, India
  • Designed and maintained ETL pipelines integrating 10+ internal data sources, reducing data processing latency by 25%.
  • Developed and optimized relational and NoSQL database schemas supporting analytics workloads for a team of 20+ engineers.
  • Achieved 99.8% data consistency in quarterly audits through robust governance and pipeline validation frameworks.
  • Delivered actionable dashboards using Power BI, collaborating cross-functionally with data science and product teams.
Broker Network Feb 2022 – May 2022
Assistant Data Engineer · Mumbai, India
  • Supported design and maintenance of data pipelines for client and broker data across reporting and analytics systems.
  • Reduced data errors by 20% through automated validation scripts on large broker datasets.
// projects

Selected AI Work

End-to-end systems built with research rigor and production mindset.

🤟
Sign Language to Text Converter
End-to-end real-time pipeline combining MediaPipe hand landmarks and a fine-tuned NLP classifier to convert American Sign Language gestures to text. Deployed as a Flask web app.
92% accuracy 26-class ASL Real-time
MediaPipeNLPOpenCVFlaskPython
📄
RAG Document Q&A Tool
Open-source document intelligence tool enabling semantic search across 100+ page PDFs with sub-2s response times. Built with LangChain and FAISS vector store, published on GitHub.
<2s response 100+ page PDFs Open-source
LangChainFAISSRAGLLM
🐄
Image Enhancement for Object Detection
Autonomous image pre-processing framework using CLAHE and illumination normalization for YOLOv8. Validated on cattle detection in low-light conditions. Research publication submitted.
+18% precision Peer-reviewed
YOLOv8CLAHEComputer VisionPython
🩻
Adaptive LBP for Kidney CT Analysis
Novel Adaptive Local Binary Pattern feature descriptor for detecting kidney abnormalities in CT scans, combined with an ensemble ML classifier. Supports early-stage clinical screening workflows.
Competitive accuracy Clinical-grade
Medical ImagingEnsemble MLFeature Engineering
☁️
AI Chatbot for Automated Booking (AWS)
Serverless chatbot on AWS automating end-to-end booking workflows with DynamoDB session persistence and SNS notifications. Reduced manual booking effort significantly.
~60% less manual effort Serverless
AWS LambdaLexStep FunctionsDynamoDB
🚗
Real-Time Vehicle & Pedestrian Detection
YOLOv8-based real-time detection system optimized for urban traffic scenarios, running at 28 FPS on edge hardware — suitable for live traffic monitoring deployment.
28 FPS edge Real-time
YOLOv8Deep LearningEdge AIOpenCV
// research

Publications & Research

Peer-reviewed research at the intersection of computer vision and applied AI.

📑
Conference Paper · Under Review
Autonomous Image Pre-Processing Strategies for Improved Object Detection Accuracy in Challenging Lighting Environments
Proposes an adaptive image enhancement pipeline combining CLAHE and illumination normalization to significantly improve YOLOv8 detection precision on low-light datasets. Demonstrated ~18% precision improvement on a cattle detection benchmark. Co-authored during Research Assistantship at the University of Alberta, affiliated with the Alberta Machine Intelligence Institute (Amii). Submitted to peer-reviewed conference venue (2024).
// education

Academic Background

Strong research foundation with specialization in AI and machine intelligence.

MSc — Computing Science (AI Specialization)
University of Alberta · Edmonton, AB
Sep 2023 – Jun 2025
Specialization: Multimedia and Artificial Intelligence. Research focus on computer vision and generative AI systems.

Affiliation: Alberta Machine Intelligence Institute (Amii) ecosystem — Canada's leading AI research hub.

Roles: Research Assistant (Computer Vision), Teaching Assistant (Software Quality)
BEng — Computer Engineering
Mumbai University · Mumbai, India
Aug 2018 – Jun 2022
Graduated with Distinction.

Final-year project on deep learning-based vehicle and pedestrian detection systems — laying the foundation for subsequent research in computer vision at the graduate level.

Early Industry: Data Engineering roles at Bosch and Broker Network during and after graduation.
// impact

By the Numbers

Quantified outcomes across research, industry, and open-source work.

92%
ASL Classification Accuracy
Real-time hand gesture recognition across 26 ASL characters
35%
Faster Query Resolution
LLM chatbot deployment reduced customer resolution time for 3 enterprise clients
18%
Detection Precision Gain
YOLOv8 improvement via CLAHE image enhancement pipeline (peer-reviewed)
20%
User Satisfaction Uplift
Chatbot A/B testing improvement via prompt engineering & fine-tuning
40%
Reduced Manual Review
Automated QA pipelines flagging model drift and compliance risks
30%
Faster Model Evaluation
Python automation across 3 dataset configurations at UAlberta
99.8%
Data Consistency
Quarterly audit score on Bosch data pipelines spanning 10+ sources
28 FPS
Edge Inference Speed
Real-time vehicle/pedestrian detection optimized for edge hardware
5+
Open-Source AI Projects
Documented AI/ML projects on GitHub including RAG tools and CV pipelines
80+
Students Mentored
Final-year CS students taught software quality; +15% avg lab score improvement
3
Enterprise Integrations
Chatbot deployments integrated with CRM and backend APIs under PIPEDA
1
Peer-Reviewed Paper
Co-authored research on image enhancement for object detection — submitted 2024
// contact

Let's Build Something
Impactful

Open to AI/ML engineering roles at top-tier tech companies. Authorized to work in Canada · Open to relocation.

✉️
Email
faisalsayed619@outlook.com
💼
LinkedIn
faisal-zafar-sayed
🐙
GitHub
github.com/faisalsayed
📱
Phone
+1 825-888-6786