💬 Ask me about Backend Engineering, LLM Systems, Machine Learning, Time Series, Data Engineering, Distributed Systems, and AI Infrastructure
I'm an AI Engineer and Backend Engineer focused on building production-grade AI systems, scalable backend architectures, and data pipelines.
🚀 Building LLM-powered systems (RAG, Agents, Structured Generation) ⚙️ Designing high-throughput backend & microservices architectures 📊 Working on data engineering pipelines (ETL, streaming, warehousing) 🧠 Deep into LLM fine-tuning (LoRA/QLoRA), inference optimization (vLLM) ☁️ Deploying systems on AWS & GCP with CI/CD pipelines 🧪 Exploring AI infra (GPU systems, local LLM clusters, distributed training)
| Area | Technologies |
|---|---|
| LLMs | GPT, Claude, Gemini, LLaMA, DeepSeek, Qwen, Mistral |
| RAG & Agents | LangChain, LlamaIndex, CrewAI |
| Fine-Tuning | LoRA, QLoRA, RLHF, DPO |
| Inference | HuggingFace, vLLM, OpenRouter |
| Vector DBs | FAISS, Weaviate, ChromaDB |
| Area | Technologies |
|---|---|
| Frameworks | Node.js, Express.js, NestJS, FastAPI (async) |
| APIs | REST, GraphQL, gRPC, WebSockets |
| Architecture | Microservices, Event-Driven |
| Messaging | Kafka, RabbitMQ, Celery |
| Area | Technologies |
|---|---|
| Pipelines | ETL / ELT, Batch + Streaming, Incremental + Backfill |
| Orchestration | Apache Airflow, Dagster |
| Warehousing | BigQuery, Snowflake, DuckDB |
| Architecture | Data Lakes, Schema Evolution |
| Type | Technologies |
|---|---|
| Relational | PostgreSQL |
| Document | MongoDB, DynamoDB |
| Wide-Column | Cassandra, ScyllaDB |
| Graph | Neo4j |
| Cache | Redis, Memcached |
| Platform | Services |
|---|---|
| AWS | EC2, S3, Lambda, ECR, ECS, SQS, RDS, CloudWatch, Bedrock, SageMaker |
| GCP | Cloud Run, GCS |
| Containers | Docker, Kubernetes |
| CI/CD | GitHub Actions, Jenkins, GitLab CI |
| Observability | Prometheus, Grafana, OpenTelemetry |
- RAG-based recruiter assistant with semantic retrieval + structured outputs
- CV parsing pipeline using LLMs + hybrid NLP + OCR (Arabic → English)
- Fine-tuned LLMs using LoRA / QLoRA on domain datasets
- Multi-model inference pipelines (Qwen, Mistral, LLaMA)
- High-throughput microservices using FastAPI + Node.js
- Async job processing with Celery + RabbitMQ
- Real-time systems with Redis + WebSockets
- Designed fault-tolerant AI pipelines
- Built end-to-end ingestion pipelines (batch + streaming)
- Designed data warehouses (BigQuery / Snowflake / DuckDB)
- Implemented incremental + backfill pipelines
- Deployed GPU inference systems using vLLM on AWS (G5/L4 instances)
- Built serverless pipelines with AWS Lambda (high concurrency)
- Containerized services using Docker + Kubernetes
- Designed CI/CD pipelines for AI deployments
- Monitoring + logging with CloudWatch + Prometheus + Grafana
🧠 Heavy focus on AI/ML (Jupyter + Python dominance) ⚙️ Backend work visible via system projects (not just commits) 📊 Lower commit count → indicates deep work vs shallow commits 🚀 Strong alignment with research + production AI systems
Build systems, not scripts. Optimize for scale, latency, and reliability. Prefer simple architectures over unnecessary complexity. Focus on real-world impact, not toy projects.
Give a ⭐ to repositories and feel free to connect!