About Us

Talent Match Africa connects experienced professionals in Africa (just like you!) with international employers from around the world (places like the USA, Australia, the UK, India, and Europe). Since 2016, we’ve been providing employability services to professionals across technology, finance, digital marketing, graphic design, and more. We currently have several permanent roles available at our offices in Johannesburg (Sandton), Durban (Umhlanga) and Nairobi (Kenya), working for incredible international companies.

About the Role

We are recruiting an LLM Developer on behalf of our client. The role of an LLM Developer is to own the model layer that decides whether an AI agent’s next action is safe. This includes intent classification on tool calls, prompt-injection and data-exfiltration detection, and risk scoring, all running inline within a latency budget measured in milliseconds, not seconds. This is applied LLM engineering under real production constraints, where rigorous evaluation matters more than demos and predictable latency is just as important as accuracy. You will work full-time as a dedicated member of our client’s engineering team.

Key Responsibilities

• Design and ship the inline guardrail and intent-classification pipeline.
• Build red-team and regression evaluation harnesses for adversarial prompts.
• Fine-tune and distil small models to meet inline latency targets.
• Build retrieval systems over policy, schema and entitlement data.
• Instrument model decisions so that every verdict is auditable and explainable.
• Collaborate with the gateway team on running inference within the enforcement path.
• Balance accuracy, latency, batching and throughput trade-offs in production.

Requirements

• Strong Python skills and comfort working in a production codebase, not just notebooks.
• Hands-on transformer fine-tuning experience, including LoRA / PEFT, distillation and quantisation.
• Ability to present a model you trained yourself, with before and after accuracy and latency results.
• Rigour around evaluation: you build the harness before you trust the number.
• Model serving experience with vLLM, TensorRT-LLM or equivalent, including latency, batching and throughput trade-offs.
• Fluency with frontier model APIs and tool-use / MCP integration patterns.
• Fluent written and spoken English.
• Advantageous: security-adjacent ML or adversarial robustness experience.
• Advantageous: real-time classification or risk-scoring experience (e.g. fraud detection in payments, banking or telecoms).
• Advantageous: experience with retrieval systems over structured or policy data.
• Advantageous: published research, open-source contributions or competition results in NLP or ML.

To Be Successful, You’ll Need To:

• Work on-site from our Nairobi, Kenya office.
• Be willing to work non-traditional local office hours, aligned with our client’s engineering team.
• Be excited about a global opportunity with an international company.
• Have excellent written and verbal communication skills in English.
• Be detail-oriented and evidence-driven, with a strong focus on measurable results.
• Take ownership of complex technical problems from design through to production.
• Be honest, ethical, transparent, and team-oriented.
• Collaborate effectively with engineering teams across different time zones.
• Be passionate about AI safety, security and building trustworthy systems.
• Thrive in a fast-paced, innovative environment while working with international teams.

We make hiring decisions based on your experience, skills and passion, as well as how you can enhance our clients’ businesses and their culture. When you apply, please tell us the pronouns you use and any reasonable adjustments you may need during the interview process.

AI experience isn't required for every role, but if you use AI to complement your work, we'd love to know how! Want to build your AI skills? Complete the free Claude courses through Anthropic Academy: Anthropic Academy

• Design and ship the inline guardrail and intent-classification pipeline.
• Build red-team and regression evaluation harnesses for adversarial prompts.
• Fine-tune and distil small models to meet inline latency targets.
• Build retrieval systems over policy, schema and entitlement data.
• Instrument model decisions so that every verdict is auditable and explainable.
• Collaborate with the gateway team on running inference within the enforcement path.
• Balance accuracy, latency, batching and throughput trade-offs in production.

• Strong Python skills
• Hands-on transformer fine-tuning experience, including LoRA / PEFT, distillation and quantisation
• Ability to present a model you trained yourself, with before and after accuracy and latency results
• Rigour around evaluation
• Model serving experience with vLLM, TensorRT-LLM or equivalent
• Fluency with frontier model APIs and tool-use / MCP integration patterns
• Fluent written and spoken English
• security-adjacent ML or adversarial robustness experience
• real-time classification or risk-scoring experience
• experience with retrieval systems over structured or policy data
• published research, open-source contributions or competition results in NLP or ML

• Work on-site from our Nairobi, Kenya office.
• Willing to work non-traditional local office hours, aligned with our client’s engineering team.
• Excited about a global opportunity with an international company.
• Excellent written and verbal communication skills in English.
• Detail-oriented and evidence-driven, with a strong focus on measurable results.
• Take ownership of complex technical problems from design through to production.
• Honest, ethical, transparent, and team-oriented.
• Collaborate effectively with engineering teams across different time zones.
• Passionate about AI safety, security and building trustworthy systems.
• Thrive in a fast-paced, innovative environment while working with international teams.