Abubakar I.

Principal AI Engineer with 7+ years of experience spanning machine learning, deep learning, and now agentic systems and LLM-driven products. I bring: Deep expertise building agentic pipelines and multi-agent systems — from self-verifying RL-tuned harnesses that autonomously evaluate and promote outputs, to fleets of coding agents operating closed-loop in parallel Hands-on experience with RAG architectures, knowledge graph construction, and hallucination-mitigation for compliance and enterprise use cases (e.g., agentic RAG assistant for banking clients) Built real-time, full-duplex voice AI systems (ASR, reasoning, synthesis under conversational latency) with barge-in and agentic screen control — voice-to-action, not just voice-to-text Strong foundation in time series forecasting and demand planning (7+ years), including large-scale SKU-level forecasting (~18,000 items) for retail, and a 1st-place Kaggle finish (30,000+ competitors) in retail forecasting Proficient across Python, SQL, PyTorch, and modern agent tooling (MCP, custom skills/tooling ecosystems), with production experience on Azure and MLOps practices (MLflow, Airflow) Track record designing reusable tools and infrastructure that let lean teams operate at multiples of their size I specialize in taking ambiguous, high-friction problems , whether forecasting, conversational AI, or full agentic workflows , and turning them into production systems that run reliably with minimal human oversight. Let's talk about how agentic AI or LLM-driven automation can move your business forward.