Research and Publication
AI-Based Energy and Water Optimization for Sustainable Agriculture
Multi-model intelligent architecture for irrigation scheduling and energy optimization in smart agriculture systems.
Research RepositoryMontassar Nawara
AI Systems Engineer
Engineering intelligent systems across Edge AI, Distributed Backend and Autonomous Agents
Edge Artificial IntelligenceAI Agents and Autonomous SystemsHigh Performance Backend EngineeringDistributed Systems
Engineering Reliable Intelligent Systems
I build production-minded software where backend performance, AI decision layers, and hardware constraints are treated as one unified system. My work focuses on real-time behavior, measurable outcomes, and architecture that scales under pressure.
From Go APIs optimized for low-latency throughput to embedded ML inference deployed on ESP32, I approach engineering as a systems discipline: performance, reliability, and intelligence must evolve together, not in isolation.
Edge AI
Lightweight inference and DSP pipelines deployed on constrained devices.
Distributed Systems
High-throughput backend services with secure APIs and in-memory optimization.
Autonomous Agents
Agent pipelines combining reasoning, orchestration, and physical action.
Selected Systems
Khayyata Platform
Scalable full-stack marketplace connecting customers with local artisans through a Go backend and React frontend.
Deployed cloud infrastructure with low-latency delivery and production-ready API integration.
GoReactAWSCloudflareREST APIs
View repositoryGOB Agent
Embedded AI voice multi-agent pipeline combining STT, intent classification, LLM inference, and TTS on ESP32.
Designed offline-resilient command workflows and near real-time control for physical devices.
ESP32PythonLLM AgentsReactCapacitorAndroid
View repositoryKINDR NLP
Real-time toxicity detection and rewriting pipeline with ML classification and LLM rewriting in one API call.
Integrated into a Chrome extension and React dashboard with live monitoring and analytics.
PythonFastAPIDetoxifyGroq/LlamaReactTypeScript
View repositoryAggression Detection AI
Audio classification pipeline from DSP feature extraction to embedded inference for verbal aggression detection.
Achieved around 82% F1-score and optimized inference for ESP32 memory and latency constraints.
scikit-learnAudio DSPMFCCESP32REST APIs
View repositoryCore Expertise
Backend and Systems
GoFastAPIREST APIsHazelcastDistributed SystemsDocker
AI and ML
scikit-learnTensorFlow/KerasLLMsNLPAudio DSPMFCCComputer Vision
Web and Frontend
ReactTypeScriptTailwind CSSCapacitorChrome Extension
Embedded and IoT
ESP32ArduinoGPIOSensorsReal-Time Systems
Cloud and Workflow
AWSCloudflareLinuxGit/GitHubAgile/ScrumUML
Engineering Experience
Backend Development Intern
DotCom - Nabeul, Tunisia
- Built high-performance Go REST APIs backed by Hazelcast with response times under 5 ms.
- Implemented API key validation and endpoint-level protection for secure production access.
- Containerized the service stack with Docker and benchmarked concurrent load scenarios.
GoHazelcastDockerfasthttpGob Serialization
AI and Embedded Systems Research Intern
ENSI Research Laboratory - Tunis, Tunisia
- Designed an end-to-end audio ML pipeline with around 82% F1-score for aggression detection.
- Built REST communication between inference services and embedded capture systems.
- Optimized model inference for ESP32 memory and latency constraints.
Pythonscikit-learnAudio DSPMFCCESP32REST APIs
Academic Formation
High-signal academic timeline. Open a card for the full orientation.
AI Systems Engineering Track - ENSI
Engineering Degree in Computer Science, Manouba, Tunisia
2024 - PresentDetails
AI Systems Engineering Track - ENSI
Engineering Degree in Computer Science, Manouba, Tunisia
Core Orientation
- Systems-oriented engineering formation centered on real-time computing, software architecture, and algorithmic rigor.
- Applied focus on intelligent systems: edge inference constraints, backend performance, and AI integration in production workflows.
- Project-driven execution across embedded hardware, distributed services, and ML pipelines with measurable latency and reliability targets.
Preparatory Engineering Cycle
Preparatory Institute of Engineering of Nabeul, Nabeul, Tunisia
2021 - 2024Details
Preparatory Engineering Cycle
Preparatory Institute of Engineering of Nabeul, Nabeul, Tunisia
Core Orientation
- Intensive scientific foundation in mathematics, physics, and analytical problem solving.
- Quantitative base applied to optimization, signal processing, and machine learning system design.
Technical Learning Path
Continuous engineering learning across AI systems, production ML, scalable infrastructure, and agentic architectures.
Applied AI and ML Systems
Machine Learning FundamentalsAI FundamentalsPractical NLP and model-serving workflowsDetails
Applied AI and ML Systems
Machine Learning FundamentalsAI FundamentalsPractical NLP and model-serving workflows
From model understanding to deployment-aware engineering decisions.
MLOps and Production Readiness
MLOps FundamentalsEvaluation and reproducibilityML lifecycle practicesDetails
MLOps and Production Readiness
MLOps FundamentalsEvaluation and reproducibilityML lifecycle practices
Shipping maintainable AI services instead of notebook-only prototypes.
Infrastructure and Platform Engineering
Containerization and VirtualizationCloud deployment patternsScalable service operationsDetails
Infrastructure and Platform Engineering
Containerization and VirtualizationCloud deployment patternsScalable service operations
Runtime reliability, portability, and operational efficiency.
LLM and Agentic Architectures
RAG with OpenAI (Azure)Retrieval + generation integrationTask-oriented AI workflowsDetails
LLM and Agentic Architectures
RAG with OpenAI (Azure)Retrieval + generation integrationTask-oriented AI workflows
Controllable LLM systems connected to backend infrastructure.
Verified
Certifications and Attestations
Selected credentials stay on the homepage, with the full archive on a dedicated page.
Featured certifications