Hello, I'm
A CS graduate building intelligent systems at the intersection of Generative AI, cloud-native MLOps, and computer vision. Former AI Engineer at Eighty4Digital. Currently finishing my FYP, FinAI, an AI-powered financial advisory platform for Pakistan.
I'm Muhammad Hasnain Akhtar, a Computer Science graduate from FAST NUCES Islamabad (2026) specializing in AI/ML, Generative AI, and cloud-native systems. I build the full pipeline: from RAG chatbots and computer vision models all the way to containerized Kubernetes deployments and GitOps-driven CI/CD.
At Eighty4Digital, I shipped production LLM chatbots with LangChain, built AI-driven automation that cut processing time by 40%, and designed DevOps pipelines with SAST/SCA security scanning, ArgoCD GitOps, and full Prometheus + Grafana observability. I also TA'd three courses at FAST, mentoring 200+ students across AI, Machine Learning, and Android development.
Gold Medalist, Rector's List of Honors, Fall 2025. Dean's List twice. AWS Certified Cloud Practitioner. Currently open to full-time AI/MLOps roles.
From production LLM chatbots to DevSecOps pipelines and computer vision research — 10 projects across AI/ML, DevOps/MLOps, and Computer Vision.
Intelligent financial advisory platform for the Pakistani market. ML models for loan eligibility, credit scoring, and financial risk prediction. Leading backend architecture as Final Year Project.
Production MLOps layer on the FinAI platform. DVC for dataset and model artifact versioning, MLflow for experiment tracking and model registry, GitHub Actions CI/CD for automated retraining, validation, and deployment.
Fully serverless order pipeline on AWS with zero infrastructure management. API Gateway + Lambda (Python 3.12), SQS FIFO for exactly-once delivery, SNS for real-time notifications, DynamoDB for storage, least-privilege IAM and CloudWatch monitoring.
Production-grade DevSecOps pipeline with security baked into every stage. Jenkins on EC2 (t2.large), SAST via SonarQube, SCA via OWASP Dependency-Check, image scanning via Trivy, deployed to Amazon EKS via ArgoCD (GitOps). Full Prometheus + Grafana observability stack.
End-to-end IaC + container orchestration pipeline. Terraform provisions a full VPC and EC2 instance; Ansible handles server configuration. Containerized React + Node.js microservices deployed to MicroK8s. GitHub Actions CI triggers ArgoCD GitOps redeploy on every push.
Novel DBSCAN algorithm for detecting small objects (humans, rocks) in LiDAR point clouds. Achieved ~17% accuracy improvement over baseline, a first-of-its-kind implementation based on research paper study.
Improved pedestrian crossing prediction on the JAAD dataset. Fused SMPL body models + CLIP with RGB frames, optical flow, and pose data in a four-component fusion model achieving higher accuracy than single-modality baselines.
Unsupervised anomaly detection system for network traffic. Implemented Isolation Forest, LOF, and DBSCAN with feature engineering, PCA dimensionality reduction, and t-SNE visualization for interpretability.
Intelligent racing agent for the TORCS simulator using reinforcement and supervised learning. Neural network architecture with Llama models for decision-making. Q-learning and policy gradient methods for race strategy.
Comprehensive fuel management system handling inventory, orders, and transactions. Complex stored procedures for automated reporting with role-based access control across multiple user tiers.
From Python and PyTorch to Terraform, ArgoCD, and LangChain: full-stack AI and DevOps.
Industry experience shipping production AI systems, plus 11 months mentoring students across three university courses.
Whether you have a role in mind, a project idea, or just want to connect — my inbox is open. I'd love to hear from you.