Generative AI Architect & Systems Engineer and Founder of Vedaris and Dueolert. Vedaris is a specialized Cloud & AI consultancy focused on production-grade cloud and GenAI systems. Dueolert is a financial management SaaS built around intelligent email processing, recurring payment detection, subscription tracking, renewal monitoring, price-change analysis, and persistent financial intelligence. Experienced in architecting multi-agent systems, RAG pipelines, agentic workflows, semantic retrieval, and scalable AI applications on AWS. Strong engineering foundation from Harvard’s CS50 AI, complemented by becoming the youngest to pass the AWS Certified Generative AI Developer Professional and 60+ industry certifications. Proven expertise across Amazon Bedrock, foundation models, embeddings, vector search, AI agents, guardrails, serverless and cloud-native AI deployment, reinforced by all 10 AWS Demonstrated credentials and 9 AWS SimuLearn pathways totaling 194 hours of hands-on training and live-console labs. Skilled in taking generative AI systems from architecture and retrieval design through orchestration, security, observability, deployment, and production scaling. Deep systems engineering knowledge spanning XNU kernel internals, macOS boot architecture, IOKit, Mach-O, UEFI, AArch64, virtual memory, kernel extensions, device drivers, graphics, firmware, and hardware-software integration. Developed a standalone UEFI Mach-O execution engine targeting Apple Silicon guest binaries on Intel hardware, incorporating structural validation, AArch64 interpretation, guest memory translation, boot verification, and SPTM/TXM boot contracts. Engineered Intel Integrated Sensor Hub communication, early-boot hardware initialization, display integration, and touchscreen coordinate transformation. Developed the first i915-on-XNU port to drive real Intel hardware, bringing Metal and Vulkan to Meteor Lake Xe-LPG on a Dell Latitude 5350 under macOS 15 through an i915-derived kernel driver on LinuxKPI, Mesa ANV, and a custom Metal integration layer. Implemented GuC submission using Linux Meteor Lake firmware and achieved a GPU-composited desktop, while resolving a LinuxKPI lost wake-up behind apparent GuC timeouts and stolen-memory allocation conflicts with the firmware console image. Experienced in Metal-to-Vulkan and AIR-to-SPIR-V translation, Vulkan compute, GPU rendering, framebuffer arbitration, and display control. Combines generative AI and cloud architecture with deep XNU kernel, graphics, firmware, and systems engineering expertise across the computing stack.
Pinned Loading
-
Meteor-Lake-Xe-LPG-Metal-Graphics
Meteor-Lake-Xe-LPG-Metal-Graphics PublicNative hardware graphics acceleration and Metal/Vulkan translation layer for Intel Meteor Lake Xe-LPG GPUs on macOS Tahoe & Sequoia.
-
KishoreKumarRVCModel
KishoreKumarRVCModel Public🎙️ KishoreKumarRVCModel — A legacy-grade voice model engineered by Ved Prajapati to capture the emotional, tonal, and phonetic essence of Kishore Kumar. Trained on 410+ curated clips with full meta…
-
-
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.






