Brando K.

⭐ CEO Intelligentia ⭐ 8+ Years of AI Experience ⭐ Expert for Agentic, Generative AI, AWS ⭐ Experience across the full AI lifecycle; from leading AI research and development teams, to designing AI products, systems, and infrastructure, to advising CEOs on AI strategy. Today, I help companies harness the full potential of agentic AI. 📌 SELECTED WORK: 🔵 Document AI Processing (pdf/tables/audio/images) - Lead AI Engineer at FileAI, Singapore Led AI Engineering Department for two years at a platform processing close to half a billion documents a year across the US, Asia, the EU and India, reporting to the CEO and owning the AI platform end to end. - Designed and shipped FileAI’s customer-facing AI assistant for automating knowledge work over a company’s own documents. Searches the user’s files, reasons over what it finds, answers with grounded citations, and generates polished PDF reports. - Generative AI OCR: researched and shipped a VLM-based OCR pipeline capturing not just raw text but document context and structure. State-of-the-art production accuracy where traditional OCR fails: Thai and Japanese scripts, long tables, checkboxes, stamps. - Agentic form filling: end-to-end system mapping extracted document content into user-defined templates, turning unstructured documents into clean structured records. Solved hard sub-problems around retrieval, multi-page table parsing, and scaling to 1000+ page documents. (LangGraph, LangChain, DeepAgents, OpenAI, Gemini, Anthropic, AWS Bedrock, AWS AgentCore, Terraform, PostgreSQL, LangSmith.) 🔵 PrivASR AI Product, fully private speech recognition Developed an End2End speech recognition solution with speaker diarization at 10x realtime for under $0.05 per hour, one-click deployable into a client's own AWS account with Terraform, at near-zero cost when idle. Whisper large-v3 on CTranslate2, pyannote voice activity detection, NeMo diarization. (Python, PyTorch, Docker, EC2, ECS) 🟠 Enterprise-grade AWS Data Mesh for a Global Reinsurance Leader Designed and deployed a multi-region multi-environment AWS Lake Formation Data Mesh solution in pure Terraform. Data Mesh operated over multiple AWS accounts representing data producers, consumers, and central data governance. Cut secure data provisioning from weeks to minutes, unblocking the company's AI programme. (AWS, Lake Formation, Terraform) 🔵 Organisation-wide AI assistant rollout for a client in Healthcare Took a company from zero to AI enabled across every team: privacy and compliance groundwork before any real data was connected, custom MCP servers wiring Claude 3P into internal systems, and AI training for each department. Teams reported doing in hours what previously took weeks. (Claude, MCP, Python, AWS) ⚫ Research Paper Publications in Peer-reviewed Journals - One-shot lip-based biometric authentication. A siamese model tying lip-movement biometrics to the spoken phrase, closing a replay-attack hole prior one-shot systems left open. 3.23% false acceptance, 3.86% false rejection on an open-set test. ISSN 0262-8856 - Automatic vision-based parking slot detection. Finds parking slots from camera images with no manual annotation, then classifies occupancy. Locates 43 of 44 slots on the PKLot UFPR05 camera with one false positive, AUC above 0.99 across every split tried. ISSN 0957-4174 🟠 Private SageMaker Studio domain, AI compute in 1-click Deployed an AWS GPU environment for a team of data scientists with no internet egress, Identity Center authentication, customer-managed encrypted EFS and private per-user storage, so no notebook, dataset or model artifact leaves the account. Terraform end to end, so a new environment comes up from one apply. (AWS SageMaker, Terraform, IAM Identity Center, EFS) ⚫ Transformer paper, 250+ Github Stars Open-source implementation of groundbreaking Vaswani, et al. "Attention is all you need." paper, 2017. 🔵 Private RAG chatbots on AWS Bedrock, for Tellofy and ComeToTheClouds Indexed website content and internal knowledge bases into support chatbots answering customer questions directly, from local proof of concept to production on AWS. (AWS Bedrock, EC2, S3, Lambda, CodePipeline) 📌 HOW I WORK: I work alongside a dozen of my own AI Agents and tools for coding, research, automated testing, deployment, etc. My agents and I can replace teams of people if the project fit is right. Before starting the engagement I want to make sure you are building the correct thing and confirm a positive ROI. I push to understand the business use case so I can emphasize on important considerations you won't think of by yourself. I will tell you when I think you are taking the wrong approach or If I am not the right person for the job. 📌 WHERE I AM NOT THE RIGHT FIT: Simple AI Wrappers, no-code automation, vibe coding - someone cheaper will do it better. I am worth the rate when the hard part is the AI / Privacy / Scale / Strategy or the infrastructure under it.