Our mission is to accelerate the adoption of human-based computation.
In other words, our goal is to facilitate the integration of human judgement into computing processes. This is especially relevant in the context of Machine Learning.
As Machine Learning gets democratized, an increasing number of automation opportunities will arise within all industries. Automation starts with routine work. This type of knowledge work tends to be the most predictable and burdensome of all, and makes it the perfect target.
Achieving full automation of a routine process is an extremely hard — if not impossible — feat in most cases. Regardless, it's a long transition that must involve both manual and automated work. And once accomplished, one should still keep a human involved for monitoring — and by extension, corrective — purposes. These setups are generally known as Human in the Loop.
For all these reasons, those who want to pursue automation opportunities within their processes need to structure them as human-based computation ones.
If you're here, that might mean you're working in an area that's comprised of one or more routine manual processes. Examples of such processes are: content moderation, data QA, identity verification or fraud. You may even be working on applying ML to a particular problem to automate it, and are relying on data annotation and/or review processes.
When a company needs to stand up one such process, they have to resort to building their own internal solution that will tailor to their needs. This is especially true as processes increase in complexity, consequently making them more unique.
However, companies rarely have the resources to adequately invest in internal solutions, thereby ending up with poor solutions plagued with tech debt, which are expensive to build.
The reality of this approach is that it makes no one happy: Product wants to minimize investment in non-user facing value, Engineering doesn't want to build "boring" Ops tooling, and Operations ends up being stuck with tooling that falls short of meeting their requirements and expectations. This is frustrating for all parties, but it shouldn't be that way.
Human Lambdas aims to empower anyone who needs to run a manual process through these core principles:
Create. Build the right tooling independently, regardless of their technical abilities or their use case.
Operate. Function efficiently at any scale, whether it is 1 or 1000 people concurrently manning the process.
Understand. Dispose of all information available to drive performance and reliability, in particular productivity, quality and latency.
We do this by building robust infrastructure that is flexible, customizable, interoperable, collaborative and transparent.
By providing reliable infrastructure that lives up to these pillars we will allow teams to easily establish human-based computation processes and help them unlock automation opportunities.