Jason Thomas Ferrell
For leaders deciding what to build, buy, fund, or greenlight, Jason turns complex AI opportunities into working software, agent systems, diligence artifacts, and film production workflows.
ContractIQ: one month, 250K lines
Enterprise-grade contract intelligence with ingestion, extraction, redlines, private MCP tools, and AWS deployment architecture.
Slate in serious filmmaker workflows
Script-to-shot production structure for continuity, reels, generated scenes, and collaborative development.
Screenvox shipped in one month
A screenplay-to-table-read product with character voices, synced animatic stills, billing, and customer workflows.
Elenqa MVP in under 100 hours
Realtime voice debate assessment with recordings, grading, exports, payments, and teacher and student workflows.
Early OpenAI partner agreement
GPT-3 legal reasoning work before enterprise LLM product patterns were obvious.
Selected clients and collaborators
Work across AI filmmaking, early OpenAI productization, private equity value creation, enterprise AI, legal evidence, and high-stakes finance.
For film and media
For producers, directors, studios, agencies, and brands.
Proof-of-concept videos and commercials
under a week
Short films
one to two weeks
Feature films
six to eight weeks
For companies and investors
For CEOs, PE/VC firms, product teams, and portfolio companies.
AI prototypes
under a week
Minimum viable products
under a month
Full production systems
two to three months
Cinema
For producers, directors, studios, agencies, and brands: development footage, commercials, shorts, feature proofs, remakes, and production systems with control over script interpretation, characters, coverage, continuity, edit rhythm, music, and tone.
Directed AI sequence / 4K sequence study
An adaptation from the 1968 screenplay, using long takes and a slow tension build in the first half before pushing into an action climax in the second.
Solo AI direction, sequence assembly, multi-model production
Public-domain remake proof / Feature study
A Roger Avary collaboration taking a low-budget 1964 public-domain film and reimagining it as a Pixar-style children's adventure with modern production value.
Collaboration with Roger Avary; AI direction, shot regeneration, score-backed proof
Historical feature proof / Feature study
A Roger Avary collaboration updating a 1927 silent aviation film with action, speed, music, and spectacle calibrated for modern audiences.
Collaboration with Roger Avary; regeneration, music-driven edit, cinematic assembly
Prestige remake proof / Feature study
A Roger Avary collaboration using Cecil B. DeMille's silent biblical epic as a test for updating archival spectacle, reverence, and coverage for modern audiences.
Collaboration with Roger Avary; visual regeneration, tone matching, production-system test
Continuity and source analysis / Scene study
A Roger Avary collaboration using Francis Ford Coppola's first film as a test for modernizing older source material through a rotoscope-inspired animation style.
Collaboration with Roger Avary; scene parsing, source analysis, shot-boundary workflow
Private crime-film proof / Private study
A Roger Avary collaboration showing modern remake capability with AI for a 1976 low-budget Italian crime film; created as a targeted demonstration for Quentin Tarantino.
Collaboration with Roger Avary; AI direction, remake strategy, scene reconstruction
Commercial proof / Portrait UGC study
A vertical commercial concept testing UGC grammar, product-centered lifestyle beats, and short-form brand pacing with AI-generated production.
Concept, AI direction, vertical commercial production
Comedy proof / Portrait joke study
An explicit classic bar-joke concept used to test character comedy, punchline timing, and vertical short-form AI video.
Concept, AI direction, comedy timing, edit supervision
Political ad proof / Portrait advocacy study
A vertical political ad sample for women's reproductive rights, testing emotional direct address, urgency, and campaign pacing.
Concept, AI direction, campaign spot production
AI Systems
For CEOs, PE/VC managing directors, and product teams: assess where AI creates or threatens enterprise value, then build the prototypes, agents, document intelligence, and secure deployments needed to test and capture it.
2026
Enterprise legal intelligence
A secure contract-intelligence product that lets a company ask serious legal questions across its own agreements.
Fractional Chief AI Officer, product architect, solo implementation lead, AWS deployment architect.
PraxisIQ needed a product that could satisfy enterprise security expectations, deploy inside customer AWS accounts, and answer complex legal questions across a company's agreements.
Built in roughly one month as a production-class system: ingestion, indexing, clause and fact extraction, redline workflows, RBAC, governance surfaces, authenticated MCP tools, and an AWS marketplace deployment model.
ContractIQ shows what enterprise AI buyers need: secure workflows, grounded answers, clear controls, and a deployment model that respects customer data.
2025-2026
Private equity value creation
AI strategy analysis for private equity: where AI threatens a company, where it can expand the product line, and where existing distribution becomes an advantage.
AI strategist, technical diligence partner, product architecture advisor.
The useful question is concrete: what could an AI-native competitor build against this company, and what should this company build first because it already owns the customer channel?
Delivered analysis across potential investments and existing portfolio companies, mapping disruption risk, product opportunity, data and workflow advantages, defensibility, execution paths, and the operating assumptions behind AI value creation.
This work gives investors a clearer view of AI risk and AI upside before capital, management attention, and engineering teams get committed.
2025-2026
AI filmmaking system
An AI director's assistant for turning scripts and source films into shot language, references, continuity, reels, and generated scenes.
Founder, product architect, full-stack builder, AI director, workflow designer.
AI video becomes production technology when the director can control sequence logic: characters, locations, coverage, continuity, camera grammar, cost, sound, and iteration.
Slate combines screenplay parsing, scene automation, shot planning, provider orchestration, visual runs, continuity checks, score generation, source-motion windows, media QA, and ongoing OpenUSD scene-engine work.
For producers and directors, Slate turns AI video from prompt experiments into a controlled development and production workflow.
One-month build
Screenplay-to-table-read product
A full product for uploading a screenplay, casting character voices, generating a performed table read, and syncing it with an animatic made from generated stills.
Founder, product architect, full-stack builder, AI workflow designer.
Writers, producers, and directors need to hear a script and feel its pacing long before a full production exists. Screenvox turned a screenplay into something closer to a pitch asset: performed, visualized, and easy to review.
Built in about one month with screenplay parsing, character extraction, voice casting, dialogue and action rendering, generated looks for characters, locations, and props, still-image animatic sequencing, audio synchronization, billing, and a customer product surface.
Screenvox ultimately evolved into Slate. It proved the screenplay-to-production workflow, then exposed the business constraint: voice-generation economics made price competition hard against better-capitalized competitors, so the opportunity moved toward deeper filmmaking control.
2020-2022
Early GPT-3 legal AI lineage
Early GPT-3 work on legal evidence, factual support, contradiction, entailment, and document-grounded reasoning.
Principal data scientist, OpenAI implementation partner, legal-AI product strategist.
Legal AI has to connect claims to evidence spans, handle contradiction, and create lawyer-grade traceability across large document sets.
Agreements with both CS DISCO and OpenAI included one of OpenAI's earliest partner agreements with a revenue-share structure. The work explored GPT-3, legal entailment, Enron annotations, contradiction datasets, zero-shot experiments, DeBERTa fine-tuning, and proof-table workflows.
This is where the later agent work started: evidence-first reasoning at scale, before LLM product patterns were obvious.
2024-2025
Omnichannel voice agents
A hardware-backed voice-agent platform for realtime audio, telephony, messaging, calendar workflows, memory, and long-running orchestration.
Founder, AI architect, hardware/software/product builder.
Useful voice agents have to survive interruptions, route tools, preserve context, resume later, use phone and calendar systems, and behave like operational assistants.
Echobach evolved from a Raspberry Pi push-to-talk device into a configurable agent platform spanning OpenAI Realtime, Twilio voice, SMS, email, contacts, RAG, workflow logs, scheduler resumes, and user-defined tools.
It pushed toward the kind of assistant people actually want: one that can speak, call, schedule, research, remember, and follow through.
2023-2024
Agentic investment assistant
An early agentic market assistant connecting tool use with portfolio state, filings, news, fundamentals, analytics, forecasting, and paid workflows.
Founder, architect, product builder, quantitative workflow designer.
Investment workflows demand tool use, traceability, market data, portfolio context, risk awareness, and judgment about when language output is useful.
Narrative Equity integrated GPT-style function calling with AlphaVantage, Finnhub, SEC data, company fundamentals, news sentiment, technical indicators, portfolio analytics, optimization routines, and Stripe-based paid flows.
It translated evidence-grounded reasoning into a market workflow before agents became a default product category.
2023-2024
High-stakes finance AI
AI architecture and strategy work for language data, internal workflows, and finance contexts where model output touches real operational risk.
AI strategy advisor, architecture consultant, evaluation and workflow partner.
In high-stakes finance, retrieval quality, latency, privacy, evaluation, data rights, and domain judgment matter as much as model capability.
NDA-limited advisory work covered language data workflows involving newswires, filings, research reports, portfolio workflows, internal AI systems, and strategy selection.
The work sits at the level where AI architecture has to withstand capital, confidentiality, speed, and institutional scrutiny.
2025
Post-essay education AI
A realtime AI debate product that lets students demonstrate understanding through argument, rebuttal, evidence, and recorded performance.
Solo MVP architect and builder across product, backend, frontend, realtime voice, grading, and payments.
When students can outsource take-home essays, assessment has to move toward live reasoning, nuance, and defensible argument.
Built in under 100 hours across teacher assignment flows, student debate sessions, OpenAI Realtime conversations, recordings, transcripts, structured grading, rubric review, exports, Supabase, FastAPI, Next.js, and Stripe.
Elenqa shows how quickly a market shift can become a tested product when product judgment and implementation stay together.
Practice
This is for teams that need more than a memo or a one-off demo: a film concept, investment thesis, product roadmap, MVP, or production AI system built quickly enough to shape a real decision.
Direct AI films, commercials, and remake studies with control over script interpretation, shot language, continuity, edit, tone, and production workflow.
Assess disruption risk, product expansion, defensibility, distribution leverage, and realistic value-creation paths for targets and portfolio companies.
Turn board-level AI ambition into product strategy, model architecture, team direction, and systems that can be shipped.
Design tools, memory, state, orchestration, evaluation, recovery, cost controls, and long-running workflows.
Bring classical ML, data science, experimentation, and measurement discipline into modern LLM product development.
Film, product, and strategy inquiries
Producers and directors can move from script, remake, or campaign idea to table reads, animatics, proof-of-concept videos, and commercials in under a week, short films in one to two weeks, and feature film builds in six to eight weeks. CEOs, investors, and product teams can move from AI thesis to prototype in under a week, MVP in under a month, and full production system in two to three months.