This document explains how hackajob leverages artificial intelligence (AI) to deliver a smarter, fairer, and more efficient recruitment experience for all users — prioritising transparency, user control, and ethical practices while maintaining a human-centric approach.

This document may be revised periodically to reflect updates in our internal processes or products.

About hackajob

hackajob is an AI-native hiring platform trusted by employers globally to find and engage qualified talent. Teams source directly or work with hackajob’s AI agents, which help identify best-fit candidates from our 1M+ private community, inbound applicants, and beyond.

On hackajob, hiring criteria is captured down to the finest detail, matching people based on skills, preferences, and real capabilities rather than keywords. Every AI-driven recommendation includes an explanation of why someone matches, so hiring teams remain fully informed and in control.

Our approach is grounded in transparency. Job seekers choose how they engage, and employers receive clear, auditable insight into how matches are generated. This ensures both sides benefit from a fair, relevant, and consistent hiring process.

AI at hackajob 

hackajob intelligence 

hackajob intelligence, our AI-native matching engine, uses algorithms to match highly relevant candidate profiles to jobs. Large Language Models (LLMs) optimise hiring processes with components that are primarily focused on natural language processing (NLP) and statistical learning algorithms rather than embodied AI or related hardware. 

hackajob intelligence combines NLP models, ranking algorithms, knowledge graphs, and LLM-powered summarisation. Models are versioned, monitored, and are evaluated for fairness, drift, and performance on a defined cycle and on material change to a model or its configuration.

Where we use AI

1.  Source

Source is a tool within the platform that lets users directly source talent, shortlist candidates, and send interview requests from hackajob’s private community. Its match and search capabilities:

• Simplify time-intensive sourcing activities such as keyword searches and complex queries, allowing more focus on higher-value tasks.

• Match candidates to job opportunities accurately and fairly, based solely on their skills, experience, and preferences.

• Support informed decision-making when reviewing candidates with clear AI insights. 

Source uses AI in the following features:

🔍 Search prompt

Your top job requirements are captured and highlighted in a single prompt which updates filters. 

↗ Job description upload

LLMs help parse job descriptions, extracting essential elements like job title, location, salary, and key skills to quickly generate a search prompt and update filters.

💼 Skills & background filters
Filters are automatically generated from the essential requirements in the search prompt. They can be edited or removed using free text or voice. Each filter includes a tooltip explaining how the platform interprets that requirement when reviewing candidate profiles.

✨ AI profile summary
LLMs provide summaries on candidate profiles highlighting how well they align with job criteria based on their skills and experience, to help inform decision making.

✅ Evidencing skills & experience
Key skills and experience are extracted and interpreted from candidate profiles, then mapped to the Skills & Background filters using natural language understanding and a knowledge graph. This highlights the evidence in each profile that aligns with the hiring criteria, providing clear, explainable reasoning behind why each profile is a strong fit.

2. Archer

Archer is an automated sourcing tool within the platform that removes the need for job creation and outreach. Archer connects to a live job feed via API to automatically create jobs and update requirements. Then displays relevant roles to matched candidates and invites them to apply directly.

• Automates time-intensive sourcing tasks, allowing time to focus on reviewing inbound
  applications, rather than engaging new talent.

• Match candidates to job opportunities accurately and fairly, based solely on their skills,
  experience, and preferences, with clear AI insights explaining why they match.

• Drives more relevant inbound applications by inviting only matched candidates to apply, increasing the proportion that progress to interview.

Archer uses AI in the following features:

Skills & background filters

Archer uses all available job information to create jobs and update the filters used in Source. It automatically displays jobs to candidates who match every Skills & background filter, as well as other essential filters like location and right to work. Candidate matches are determined by extracting and interpreting skills and experience and mapping them to filters using natural language understanding and a knowledge graph.

Invitation to apply

Candidates who match every essential filter for a role are shown that role and invited to consider it. The invitation is an opportunity presented to the candidate, not a decision about them: the candidate chooses whether to apply, and applies through the employer’s own application process. No ranking, score or assessment output is provided to the employer, and no candidate is rejected or excluded from a process by Archer.

3. Thena

Thena is a tool within the platform that helps users screen high volumes of direct applications for selected roles. Thena screens applicants in hackajob via a direct ATS integration, ranking them in order of best fit with explainable recommendations. This supports users to:

• Improve application screening efficiency by displaying the strongest applicants first.

• Confidently make progression decisions with clear, explainable insights into why someone may be suitable.

• Provide a fairer and faster candidate experience by reviewing and responding to
  applicants sooner.

Thena uses AI in the following features:

Skills & background filters

Filters are generated from the hiring criteria defined by the employer. They are populated using natural language understanding and can be edited or removed using free text or voice. Each filter includes a tooltip explaining how the platform interprets the requirement when reviewing applicant profiles. Applicants are then ranked based on how closely they match these filters.

Evidencing skills & background on profiles

Key skills and experience are extracted and interpreted from candidate profiles, then mapped to the Skills & Background filters using natural language understanding and a knowledge graph. This highlights the evidence in each profile that aligns with the hiring criteria. This provides clear, explainable reasoning behind why an applicant is or isn’t a strong fit.

Thena can be switched on or off at any time, giving users full control when automated
screening is applied to a role.

User control & transparency

We understand the importance of trust when using AI-powered tools, ensuring that AI acts as a supportive assistant, rather than dictating user activity. 

AI transparency

• AI-driven features such as the profile summary are highlighted or labelled, so users know when their actions are supported by AI. 

• Explanations on filters and profiles provide clear evidence and explain why candidates are recommended.

Transparency to candidates

• Candidates are told when they are interacting with AI. Messages generated by hackajob’s AI agents identify themselves as AI-generated, and hackajob’s use of AI to match candidates to roles is described in its privacy and in this guide. 

User control

• Users are always in control. While AI provides insights and suggestions, progression decisions (interview, shortlist, or reject) are made by users, ensuring full control at every step.

• Users can fine-tune candidate matches using filters or by adjusting the search prompt at any time, to ensure results align with their preferences.

Human review

• Any candidate may request that a human review an AI-supported action taken in relation to them, rather than relying on the automated recommendation alone. The contact route for making that request is set out in hackajob’s privacy policy. Hiring decisions are made by the employer’s recruiters and hiring managers in all cases.

Ethics & bias-mitigation

We’re committed to ensuring fairness, mitigating bias, and adhering to ethical AI standards.

Fairness

Bias audits

• Regular reviews are conducted to detect and reduce biases that might unfairly favour certain groups.

Objective candidate ranking

• Candidates are ranked exclusively on objective data inputs they provide, such as skills, experience, and preferences–excluding irrelevant factors like gender or ethnicity.

• AI evaluates candidates’ profiles based on skills, experience, and preferences, prioritising fair and consistent assessments over factors like formatting or writing style.

Pre-trained models; no customer data used for training

• hackajob does not train or develop foundational AI models. Its AI systems are built on pre-trained third-party and self-hosted large language models, combined with hackajob’s own orchestration logic, natural language understanding and knowledge graph. Customer data, candidate profiles, job descriptions and ATS data are used for real-time inference and evaluation only, and are never used to train hackajob’s own or any third-party’s general-purpose models.

Safeguarding

Data protection

• Strict security measures ensure that no personally identifiable information (PII), such as gender or ethnicity is exposed to AI models, protecting candidate anonymity.

  • Diversity data is collected with explicit consent from candidates to share with employers and never inferred.

Data minimisation & isolation

• Only the minimum data required to generate match explanations or insights is processed by AI components

• Customer and candidate data is logically isolated and cannot be cross-referenced between clients.

AI interaction 

• AI tasks are performed in a secure and controlled environment, without access to sensitive user data.  Monitoring tools detect and prevent unauthorized access or data leakage.

• We do not use customer data, candidate profiles, job descriptions, or ATS data to train general-purpose models.

• Data is used only for real-time inference and evaluation.

Continuous monitoring

Evaluating fairness

• Regular monitoring and periodic human validation detect and address unintended biases. 

Performance

• The system is monitored in real-time to detect anomalies and ensure it operates within accepted thresholds, with immediate action taken to address issues and maintain reliability and fairness. 

User feedback

• We actively collect user feedback to improve AI processes, keeping recommendations unbiased and aligned with user needs.

Raising a concern

• Concerns about bias, discrimination or the behaviour of an AI-supported feature can be raised by a candidate, a client or hackajob personnel using the contact details in hackajob’s privacy policy or through the client’s account contact. Concerns are logged and given a severity classification at initial triage, which determines the escalation path and the response times that apply under hackajob’s AI Incident Management Policy. Where personal data is involved, the Data Protection Officer is engaged. Corrective actions are assigned named owners, recorded in hackajob’s AI Incident Register and tracked to closure, and a root cause analysis is carried out. Incident trends are reviewed by hackajob’s AI Ethics & Risk Committee and fed back into the risk assessment for the system concerned.  Where a concern is substantiated, containment measures are applied before remediation, which may include withdrawing the affected model version from production.

Documentation

• hackajob maintains a Model Card, a Bias Report and an Accuracy Report for the AI systems in its platform, produced and reviewed under its AI management system. These are available to clients on request, subject to a confidentiality undertaking.

Security and data protection practices

We are committed to safeguarding user data and maintaining high standards of security, privacy, and compliance across all regions in which we operate. Our approach aligns with internationally recognised frameworks and evolving global regulatory requirements.

ISO/IEC 27001:2022 Certification

hackajob operates a certified ISO/IEC 27001:2022 Information Security Management System (ISMS), covering the people, processes, and technologies that support our platform. This includes:

• Risk assessment and continuous improvement processes

• Strong access control and least-privilege governance

• Encryption of data in transit and at rest

• Secure software development practices (SSDLC)

• Regular internal and third-party audits

ISO/IEC 42001:2023 Certification

hackajob operates a certified ISO/IEC 42001 Artificial Intelligence Management System (AIMS), the international standard for the responsible governance of AI. Certification is independently audited and subject to ongoing surveillance. The management system covers:

• AI risk assessment for each AI system

• Bias and fairness assessment, with defined review cycles and re-assessment on material change to a model or its configuration

• Model documentation, version control and change management

• Monitoring of performance, drift and fairness

• Human oversight requirements and AI incident management

• Supplier assessment for AI providers

SOC 2 Type II

hackajob maintains a SOC 2 Type II report, an independent attestation of both the design and the operating effectiveness of its controls across security, availability, processing integrity, confidentiality and privacy, assessed over a defined review period. These controls include:

• Logical and physical access controls

• System and network monitoring

• Third-party/vendor risk management

• Incident detection and response

Global Privacy & Data Protection Compliance

hackajob adheres to applicable data protection and privacy laws in the regions where we operate and process data, including:

• UK GDPR and the Data Protection Act 2018

• EU GDPR 

• US state privacy laws, including CCPA/CPRA and similar enabling state-level regulations

Our privacy practices include:

• Clear allocation of Controller/Processor roles depending on the product and use case

• Data minimisation and purpose limitation

• Support for data subject/consumer rights (access, deletion, correction, opt-out)

• No use of customer or candidate data to train shared or general-purpose AI models

• Logical separation of customer data and strict internal access controls

• Secure deletion processes at the end of the relationship or upon request, in line with legal obligations

AI providers and data location

hackajob platform is hosted on Amazon Web Services in Ireland, within the European Economic Area. Some AI inference providers are located outside the EEA, principally in the United States; any such transfer is governed by a data processing agreement incorporating the EU Standard Contractual Clauses and the UK International Data Transfer Addendum, supported by a transfer impact assessment. A current list of hackajob’s sub-processors, including AI providers, is maintained at trust.hackajob.com. 

Every AI provider hackajob engages is subject to the same standing commitments: client data is submitted for real-time inference only, and no client data is used to train any model, the provider’s or hackajob’s. Where a provider offers a zero-retention configuration, hackajob operates on that basis. Each provider passes a vendor security assessment before onboarding and is reviewed at least annually, and the retention position for each provider is set out in hackajob’s data processing documentation, available to clients on request.