Looking for a job

Muhammad Anns Khan

@muhammadannskhan

AI Engineer at MeliorITS building multilingual document retrieval systems, with production ingestion for 15,152 manuals.

Pakistan

Message

What I'm looking for

I’m looking for senior AI engineering, ML platform, model-serving, or production GenAI work where I can contribute as an individual contributor.

At MeliorITS, I built document-understanding and retrieval infrastructure for a field-service AI assistant answering equipment-specific questions from technical manuals. I replaced a GraphRAG / Azure AI-based ingestion path with local encoder models and deterministic extraction, retaining page- and region-level evidence for answers.

I architected a 21-stage document pipeline and deployed an API with five worker services on an NVIDIA H100 host. It processed 15,152 documents for six companies, while hybrid retrieval reduced observed search latency from 10–45 seconds to 0.1–0.8 seconds.

At Veroke, I work on MiiQ, a multi-tenant enterprise knowledge platform. I designed a durable GPU execution broker and reworked ingestion into independently queued, resumable stages, with capacity controls and recovery behavior for failed work.

I also instrumented production observability across 131 metric families and published 12 Grafana dashboards and 32 Prometheus alert rules. For AI safety, I integrated tenant and content-policy checks into live retrieval and helped close critical and high-priority findings from a technical guardrail audit.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

MeliorITS

Oct 2026 - Present (1 month)

Built document-understanding and retrieval infrastructure for a field-service AI assistant, replacing GraphRAG with local encoder models and deterministic extraction. Architected a 21-stage validated DAG for document intake, layout analysis, OCR, and indexing, and deployed production services on NVIDIA H100 with PostgreSQL/pgvector.

Python fastAPI PyTorch Hugging Face ONNX Runtime GLiNER KeyBERT Multilingual E5 Embeddings DINOv3 Paddle OCR PyMuPDF PostgreSQL Pgvector HNSW Docker Celery RabbitMQ Redis Prometheus Grafana Azure DevOps

Current

AI Engineer

Current

Veroke

Aug 2024 - Present (2 years 3 months)

Core engineer for MiiQ, a multi-tenant enterprise knowledge platform, delivering production components across document processing, GPU execution, inference routing, orchestration, observability, resilience, and policy enforcement. Focused on failure-safe architecture and verifying critical features in live workflows.