AI engineering · Software systems
Awab Zemzemi
I’m a final-year engineering student at INSAT. I build AI agents, retrieval systems, and the Python services and infrastructure that run them.
Case studies
Selected projects
Infrastructure diagnostics
Read case studyAutonomous AIOps Agent
Investigates Kubernetes failures using LangGraph, restricted diagnostic tools, and namespace-scoped authorization.
LangGraph · FastAPI · Kubernetes · RBAC
Investigation flow - FastAPIAuthorize namespace
- LangGraphSelect a tool
- Diagnostic toolsCollect evidence
- KubernetesEnforce read-only RBAC
Knowledge retrieval
Read case studySecond Brain
Combines semantic search and knowledge-graph relationships through LangGraph to answer questions across personal notes.
LangGraph · Neo4j · ChromaDB · LangSmith
Knowledge graph schema MLOps · Streaming ML
Read case studySentinelML
Pairs MLflow-tracked Isolation Forest training with Kafka transaction processing and a registry-backed inference worker.
Kafka · Python · MLflow · Docker · Jenkins
Transaction scoring pipeline - TransactionsDataset replay
- KafkaRaw event stream
- FeaturesProcess and publish
- InferenceScore → predictions
Background
Experience
Professional experience
Bit&Books Inc.
AI Intern
- Built a PaddleOCR-VL and LLM pipeline that converts invoices and receipts into bookkeeping transactions, recovers missing fields and line items, and checks arithmetic and totals.
- Used BGE-M3 and Qdrant to retrieve Chart of Accounts candidates, then used an LLM to select the account and flag uncertain classifications.
- Deployed vLLM inference on RunPod GPUs and wired it into a Python backend with a task queue and parallel workers. Processed 1,880 document-processing items during the internship.
Python · FastAPI · PaddleOCR-VL · BGE-M3 · Qdrant · vLLM · RunPod
KIAN Technology
DevOps Intern
- Deployed Wazuh on DigitalOcean Kubernetes for security monitoring and integrated Falco DaemonSets for webhook-based runtime alerts.
- Automated infrastructure provisioning with Terraform.
Kubernetes · Wazuh · Falco · Terraform · DigitalOcean
ACTIA
Mobile & Web Development Intern
- Developed an e-bike battery monitoring app and battery-health dashboard using Flutter, Flask, and Firebase.
- Automated builds with Docker and Jenkins CI/CD.
Flutter · Python · Flask · Firebase · Docker · Jenkins
Education
National Institute of Applied Sciences and Technology (INSAT)
Engineering Degree in Computer Networks and Telecommunications
Machine Learning and Deep Learning · Software Architecture · Microservices · Cloud Computing
About
Engineering interests
I study Computer Networks and Telecommunications at INSAT and build AI agents and LLM applications. My work connects retrieval and model inference with the backend services that make them usable.
I’ve worked on document processing, GPU-hosted inference, and Kubernetes infrastructure. I’m interested in making those systems reliable and maintainable, from validating model output to tracing failures after deployment.
Contact
Let’s talk systems.
Working on an agent that needs a reliable backend? I’d love to hear about the engineering decisions behind it.