Az-har Az-har

Hi there, I'm Azhar 👋

Data Engineer specializing in Celonis Process Mining, enterprise data pipelines, and autonomous applied AI.

I design and build enterprise data platforms, process intelligence systems, and applied AI pipelines. My core focus spans:

Celonis Process Mining: Deep experience in Order-to-Cash (O2C) and Procure-to-Pay (P2P) event-log modeling, PQL analytics, Celonis Studio, and Action Engine automated workflows.
Modern Data Engineering: Architecting reliable batch, dimensional, and lakehouse pipelines using Databricks, Apache Spark, dbt, DuckDB, and PostgreSQL.
Autonomous AI & Pipeline Systems: Engineering hardware-accelerated local speech engines (Whisper Vulkan), LLM evaluation loops, and enterprise RAG validation frameworks.

Enterprise Order-to-Cash (O2C) Process Intelligence & AI Engine

Process Bottleneck Resolution: Flags delivery delays and resolves execution gaps in transactional SAP/Celonis O2C flows.
Databricks Master Pipeline: Scalable daily scoring with vectorized batch inference, environmental caching, and automated ML evaluation.
Multi-Input RAG Validation: Semantic ChromaDB retrieval cross-verifying contractual SLAs, disruption documents, and penalty exposure ($).

Autonomous 5-Agent Media & Audio Mastering Pipeline

Acoustic Perception: Hardware-accelerated Whisper large-v3 on AMD GPUs via Vulkan compute with Silero VAD neural pre-filtering.
Self-Evaluating Critic Loop: Autonomous multi-turn LLM judge auditing factual consistency, narrative flow, and ad purging.
Broadcast Mastering: Studio-grade audio normalization to EBU R128 (-16 LUFS) with ID3v2 chapter embedding.

automotive-analytics-project: Modern analytics engineering pipeline transforming raw automotive datasets into multi-tier dimensional marts using dbt, DuckDB, and SQL.
scratchpad-python: Curated standalone Python tools, 3D mathematical terminal graphics (donut.py), and CLI utilities.

_{Engineered by Azhar • Specialized in Celonis Process Mining, Data Engineering, and Applied AI}