ASAKASA SRL / [CUI pending] / Bucharest
Asakasa builds AI-native systems that observe filings, macro data, options flow, prediction markets, insider activity, and geopolitical risk, then compress them into operator-grade decisions.
AI is research and engineering leverage here. It helps build, audit, summarize, and operate the machinery. It does not replace operator judgment.
What we build
Systems built and operated internally. Selected research surfaces may be available to qualified collaborators on request.
REGIME / ALERTS / BRIEFS
Raven watches filings, macro data, options flow, insider activity, public prediction-market flow, and geopolitical risk. Its role is not to produce noise; it escalates only when the evidence clears a high-conviction bar.
POLYMARKET / CRYPTO / ROUTING
Translight is Asakasa's execution research platform for Polymarket and crypto venues. It studies routing, latency, dry-run gates, paper-shadow behavior, live canaries, and operational controls before capital is exposed.
DIRECTIONAL / TESTING / FALSIFICATION
Yutani tests fast directional signals under pre-registration, calibration, fees, slippage, and out-of-sample pressure. Its strongest lesson is discipline: a gate that cannot survive prospective evidence gets demoted.
MAKER / BASIS / INVENTORY
Hyperliquid research studies passive liquidity, funding and basis, inventory limits, HLP baselines, and hard risk envelopes. The question is not whether a model looks clever; it is whether execution economics survive contact with the venue.
Operating doctrine
01
Evidence before action
02
Regime before signal
03
Slippage before PnL
04
Kill criteria before deployment
05
Operator custody before automation
The lab does not treat signal volume as intelligence. A useful system has to know when a market is outside its edge regime, when costs dominate the idea, when a backtest is decorative, and when the correct operator action is to wait. Automation is allowed to observe, compress, test, and alert. Custody of action remains with the operator.
Signal discipline
Trust, after corroboration
Distrust, unless proven otherwise
Thesis
Asakasa studies the regimes where technology, geopolitics, liquidity, and narrative collide. The obvious trades are crowded quickly. The less obvious constraints often sit in filings, procurement calendars, supply chains, option surfaces, prediction markets, insider timing, and the friction between policy ambition and physical capacity.
The lab's edge is not the claim that a model can see the future. The edge is building systems that force every candidate signal through regime context, source quality, transaction costs, capacity, and invalidation criteria before it reaches an operator.
Raven is the public-facing intelligence layer. It monitors regime transitions, asymmetric technology themes, bottom and top conditions, black-swan transition risk, public prediction-market repricing, and insider activity when the activity is material enough to matter.
Translight, Yutani, and Hyperliquid research form the execution and falsification layer underneath that intelligence work. They ask whether a signal survives real venue mechanics: latency, fees, slippage, maker fill probability, wallet risk, dry-run gates, canaries, kill switches, and the unglamorous cost of being wrong.
AI is used as engineering and research leverage. It helps build scanners, summarize messy evidence, draft audits, generate adversarial checks, and compress broad data into readable briefs. It is not presented as a magic alpha source. The operator owns judgment; the system disciplines that judgment.
The result is a research operation that treats waiting as an active decision. A high-quality alert should be rare. A useful brief should say what changed, why it matters, what would invalidate the read, and whether the current regime permits action at all.
Asakasa exists for operators who care about asymmetric technology regimes: AI, semiconductors, defence, energy, space, cyber, prediction markets, and the political constraints around them. These regimes move through bottlenecks before they move through consensus.
Register interest in the Asakasa Intelligence Brief: market regimes, event risk, asymmetric technology themes, and high-conviction operator notes.
SubscribeDiscuss research infrastructure, execution systems, data pipelines, audits, and AI-augmented engineering operations.
CollaboratePrivate inquiries for operators who need disciplined market intelligence, risk gates, and execution research around hard constraints.
Inquire