Applied Reason

Selected work / Engineering systems

Engineering behaviour,
made visible.

Five systems spanning operational AI, media orchestration, attention design, recruitment software and physical hardware bring-up.

01 / Operational intelligence

RunTime Recruitment
Intelligence Platform

What it teachesA request becomes parallel tool activity, waits for asynchronous responses, then passes through assessment and human approval before an operational action.

RunTime Recruitment Intelligence orchestrationA request enters orchestration, branches to ATS, semantic search, candidate data and API tools, reconverges for assessment, human review and approved action. REQUESTVOICE · TELEGRAMORCHESTRATIONWORKFLOW AGENTSMCP FOUNDATION* ATSJOBS · WORKFLOWSEMANTIC SEARCHQDRANTCANDIDATE DATAPROFILES · RESUMESAPIsENRICHMENT · TOOLS RESPONSE ASSEMBLYASYNC COMPLETIONAI ASSESSMENTHUMAN REVIEWAPPROVAL GATE REQUESTAPPROVED ACTION RELEASED* MCP foundations verified; live transport trace not represented.
Asynchronous fan-out and convergence — approved pilot baseline

02 / Media orchestration

AI Video Content
Operating System

What it teachesGenerative media and audio analysis are separate asynchronous experiments. The concept does not claim an editing, rendering or publishing system.

AI video prototype jobsA source brief branches into text-to-video, image-to-video, and audio transcription prototype jobs. Unverified editing and publishing stages are crossed out. SOURCE BRIEF / MEDIAPREPARESCENES · PROMPTSMANUAL TRIGGER TEXT → VIDEOASYNC REQUESTSTATUS POLLINGIMAGE → VIDEOASYNC REQUESTSTATUS POLLINGAUDIO EXTRACTIONTRANSCRIPTIONPERSISTENCE VIDEO URLsCOLLECTEDTRANSCRIPT+ SUMMARYBASEROWEDIT · RENDER · PUBLISHNOT VERIFIEDPROTOTYPE OUTPUTS COLLECTED
Transformation lanes — verified prototype scope, not a claimed finished pipeline

03 / Attention system

Ambient
Frame

What it teachesMany signals compete for attention. A context and priority layer allows one relevant item through. This remains an explicitly unverified product and hardware concept.

Ambient Frame filtering conceptSix illustrative signal sources enter a narrowing context and priority filter. Most are suppressed and one reaches a conceptual display.CALENDARTASKSEMAILPROJECTSGOALSCONTEXTCONTEXTPRIORITY / RELEVANCEDECISIONSUPPRESSED / DEFERREDNEXTDesign review14:30ONE RELEVANT ITEMCONCEPT · HARDWARE UNVALIDATED
Progressive filtering — less information, better attention

04 / Operational platform

RunTime
ATS

What it teachesA requirement changes a connected record system: job context drives semantic matching and assessment; recruiter review releases an auditable workflow action.

RunTime ATS relationship graphJob, candidate, company, communication, assessment and workflow records are connected. A requirement triggers search, assessment, review and an approval-gated action.JOBREQUIREMENTCANDIDATEPROFILE · RESUMECOMPANYCONTEXTASSESSMENTAI + REVIEWWORKFLOWQUEUES · ACTIONSCOMMUNICATIONAPPROVAL-GATEDNEW REQUIREMENTSEMANTIC MATCHAI-ASSISTED ASSESSMENTRECRUITER REVIEWWORKFLOW ACTION
Relationship graph — one event updates a mature operational system

05 / Hardware bring-up

INMOS B004 / T400
Transputer Bring-up

What it teachesModern Linux tooling crosses a USB-to-ISA interface to bootstrap physical INMOS hardware, execute code and return a verified result.

INMOS B004 T400 Linux bring-up schematicLinux diagnostics, TDS2 server and iserver pass commands through an ARS USB2ISA bridge to an ISA B004 board and T400 transputer, then return a verified result.LINUX HOSTb004diag · t400memprobeTDS2 SERVER · ISERVERARSUSB2ISA BRIDGEINMOS IMS B004 · ISAMEMORYPATHT400PHYSICALSILICON01 HOST COMMAND02 ISA TRANSFER03 BOOTSTRAP04 EXECUTE05 VERIFIED RESULTBOOT / DATARESULTVERIFIED: C HOST TOOLS · ARS SDK INTERFACE · LINUX TDS2 SERVER · PORTED ISERVER
Low-level schematic — software → interface → hardware → execution