Why join Liquidity Labs

MM benefit comparison

Liquidity Labs are currently a two-man-show project providing liquidity on centralized crypto­currency exchanges, which means we simultaneously quote to buy and sell tokens. We trade in a way that is useful for the ecosystem. We do not speculate, do not participate in foul play such as manipulation or front-running. We try to contribute to stability and liquidity of the cryptocurrency markets so that crypto is less volatile, less risky and positions are less expensive to execute. And if we do it right, we can earn a lot of money in a highly scalable way too. By taking part of the profit from big companies and distributing them into our team, we are contributing to decentralization of the crypto­currency market too. We try to be efficient and focused and it works well so far – we consistently achieve returns around 100% p.a.

So this is our mission. We are looking for skilled and motivated people who want to take part in it.

Who do we need

We currently need these people:

  • a great Python developer with some Linux/commandline skills for work on production implementation and analysis engine. C++/Cython skills are advantage. You have to know how to use debugger and profiler, write reliable code and have some problem-solving skills. He should have some overlap into trading and understand how strategies work to some extent.
  • a talented Python quantitative analyst with good command of Python & Pandas and ideally some experience with simple statistical/ML modelling. You should know how to use pandas groupby, git, basic statistics, understand market microstructure (orderbook, trades) and some of its implications. It would be good if he was not afraid of a bit of production system implementation too.
  • and also an insightful consultant with experience in market making companies, preferably in crypto­currency space, for a few hours of consultations monthly.

The first two roles should be available for at least 20h a week. We will be working remotely (ocassional coworking is preferred if you’re around). Therefore a good degree of independence is required. We will sign a contractor agreement with monthly payment.

What do we need

We recently finished a setup allowing us to trade with around 2.5 bps fees on Binance and want to scale up the volume and leverage, which should multiply our returns. Here is where you come in. Technically this is a rough outline what you might be doing in the first 6 months depending on your specialization and interests:

  • Platform development
    • networking magic: sending some request (cancels) concurrently to multiple API IPs to lower latency
    • networking magic: hacking dns resolving
    • computing overall (beta) exposure across several instances of system and hedging the exposure on correlated futures
    • chase down bugs from our Sentry bug management system
    • optimize performance and minimize latencies
    • running the whole production platform on historical data replay/simulation
    • improve tests and documentation
  • Development / quant combination
    • add new features into EV computation (response function, trade sign autocorrelation)
    • consider spread sizes on different pairs and calculate expected profitability of possible stat arb trades
    • update quoting logic to balance usdt/busd/btc/eth base token inventory well
    • how to send orders so that we don’t exceed exchange order creation rate limits, but keep our orders up to date?
    • use taker orders to manage inventory in extreme cases
  • Quant research
    • creating grafana charts analyzing results of our trading and drawing conclusions from them
    • maybe move the grafana analysis into jupyter notebook(s)?
    • simple statistical models, eg. how orderbook imbalance affects future price?
    • combining and backtesting the above
    • implementing analytical backtesting engine for the above
    • (semi-automating) explorative analysis of our fills, eg. how did the worst fills look like, could we have prevented them based on some existing/new feature?
  • Operations
    • taking care of our 10-40 AWS servers
    • implementing simple Ansible deployment or dockerizing stuff

Results

We were pretty successful so far, just in the last 12 months we reached:

  • 28% p.a. return (more recently 42% p.a. in last 6 months, 96% p.a. in the last month)
  • Max drawdown 8% after FTX collapse. Back to all time high in two months.
  • We were aproximately market-neutral, we hedge the whole portfolio on futures to minimize the effect of market trend on our profitability.

Our offer

Compensation package consists of an agreed combination of fixed salary, personal bonus and most importantly a possibility of depositing your money into the fund. For experienced and long-term coworkers, it might be possible to get equity options as well.

Example: Full time contract, senior role, $3000 base. Up to $20k fund allocation in first 6 months, then $60k. In the last 12 months that would yield ~$7160/mo = ~155 000 Kc/mo total without taking any personal bonus into account.

We expect our profits will grow 2-10x in the next 12 months, so the salary from above example might get closer to $20 000/mo. The contract and fund allocations are reviewed yearly.

Participation: being part of tiny startup where you can double the revenues in a few months of work. Ocassional teambuildings, perhaps international later? Participation in huge upside.

Meaningful job: we don’t speculate, we reduce market slippage and volatility. This helps millions of crypto traders, even though it’s just a small difference we make. Our edge is systematic and will not disappear overnight.

We already started a small initiative benefitial to the community: affordable market data service for individuals, where we don’t aim for making money, but rather to help democratize crypto quant trading space, so that individual quants can compete with and beat big trading companies.

In the future we would like to help educate people in crypto­currency investment space. I believe that crypto is a trap for many people who don’t really understand how things work. So they loose money by frequent day-trading or by blindly following influencers shilling their bags. Those people are creating some of the inefficiencies we trade, so it would only be fair to give back.

Our tech stack

We model prices and volatility based on the actual market micro-structure. We use Python/Jupyter/Pandas and employ statistical and simple ML (GBT) techniques, but most of our alpha is in our infrastructure. We have a fully automated Python-based pretty-low-latency trading platform with spartan UI and a real-time evaluation system with charting in Grafana.

Grafana screenshot: top left shows total profitability in time, top right shows A/B test of new ML model, bottom shows profitability on individual pairs.