Khemaji Thakor

Senior Applied AI Engineer with 6+ years of experience building and shipping production AI systems and Python backend platforms. Most of my work sits at the intersection of NLP, event-driven backend infrastructure, and enterprise automation shipping OCR pipelines, RAG-powered knowledge systems, and semantic compliance engines used in live logistics and customs operations. Specialized in Python, Django, Kafka-based asynchronous architectures, NLP pipelines, OCR systems, semantic retrieval, and RAG-powered enterprise knowledge platforms.

Experienced in designing event-driven backend services, AI orchestration workflows, vector-search systems, and observability-enabled infrastructure supporting high-volume operational processing.

Senior AI/ML Engineer at Intech Creative Services (2022-07 – Present)

• Led backend architecture for three production AI platforms a logistics orchestration engine, a customs NLP compliance system, and an enterprise RAG knowledge base all built on Python/Django with Kafka handling async event processing at 50K+ events/day.
• Designed the AutoTrip shipment-planning system: geospatial clustering, vehicle-capacity optimization, and route planning reduced delivery trips by 18% and improved fleet utilization by 25% across active logistics operations.
• Built a customs compliance NLP pipeline that ran semantic similarity scoring across 50K+ cargo records daily to flag mismatches between commodity descriptions and HS codes, cutting manual review volume by ~40%.
• Deployed an RAG knowledge platform (OpenAI GPT-4 + ChromaDB) with a custom chunking and embedding pipeline, enabling staff to retrieve accurate answers from 10K+ internal documents in under 30 seconds vs.~15 minutes of manual search.
• Set up centralized observability across all backend services using Prometheus, Grafana, and Loki — reduced mean time to detect incidents from hours to minutes and maintained 99.5%+ uptime SLA.

Machine Learning Engineer at Samcom Technobrains (2020-02 – 2022-07)

• Built intelligent data-processing pipelines for logistics and compliance clients using Python, Django, PostgreSQL, and AWS (S3, EC2) supported high-volume ingestion workflows across multiple concurrent projects.
• Developed an ML text-classification pipeline for Innovation DB that auto-categorised 50K+ research records, replacing a 3-person manual tagging process and reducing categorisation time by roughly 70%.
• Engineered an OCR-based document extraction system for shipment data parsed, validated, and classified scanned manifests with 90%+ field extraction accuracy, feeding downstream compliance workflows.
• Built REST APIs and role-based access-control modules for an enterprise process management platform, supporting operational monitoring dashboards used by ~50 internal users daily.

Machine Learning Engineer at Azine Technologies (2019-07 – 2020-02)

• Developed OCR-based document parsing and identity-verification workflows for the IDMerit platform using Python, Django, and Tesseract OCR integrated with a REST API layer consumed by a mobile client.
• Built backend modules for approval workflows, audit tracking, and document lifecycle management in an enterprise identity-management system.

Bachelor of Engineering in Information Technology – Gujarat Technological University