AI Technologies

The problem

Asset management managers and private equity analysts spend significant time just reviewing company reports, understand them and producing their own reports for their investors. The process at the moment is manual and time consuming.

THE SOLUTION

We build an agent system (with a fine tuned LLM and various 'smaller' AI models for specific tasks) that produces automatically quarterly reports (with sources' references on each paragraph) on each company of the portfolio just uploading the ﬁnancial reports. The agent can be queried on specific questions and can execute a range of specific tasks on documents (and online) in real time.

The problem

At the moment drone controllers need to constantly manage the device while in action. There is a need to develop solution to control a drone by instructions ('go to the tree', 'follow a red car' etc.) instead of commands (joystick left, right etc.).

THE SOLUTION

W have developed agents able to complete certain missions in a 'game environment'. The hybrid human-reinforcement learning algorithm is able to play the videogame 30% faster than any machine alone and with less 'constant focus' from the controller.

The problem

The client is an African telecommunication company with more than 2 million users. By law the company needs to review the network traffic data and produce reports about cyber security attacks, malware, failed connections etc. Producing these reports manually takes significant time of the SOC team.

THE SOLUTION

We designed and developed an automated AI solution to detect malicious behaviours in traffic workﬂows integrating output from different hardwares and software solutions (switchers, servers, mail data software etc.). Developed a pipeline to automatically report alerts and incidents. Improved effciencies of the analysts’ team (SOC) work of 100%.

The problem

The client is an institution with both sensitive data and live AI solutions. For security reasons and by law the AI solutions (and data) need to have constant checks in place (to avoid corruption of the output and mitigate potential unwanted results).

THE SOLUTION

We designed and developed an AI governance system (policies, procedures, tests, registers etc.) to ensure the speciﬁc AI solutions are not being corrupted by cyber attacks and, if any, the corruption is mitigated by triggering a pre-planned incident response procedure.

Cybersecurity

Case studies:

The problem

THE SOLUTION

The problem

THE SOLUTION

The problem

THE SOLUTION

The problem

THE SOLUTION

How can we help you?