— CAREERS
WORK WORTH BEING RIGHT ABOUT.
A model is a claim about the world. We are looking for people who can carry that responsibility.
– WHY ALETHEIA
JUDGEMENT IS THE JOB.
Not the timesheet, not the deck, not the process around it.
Aletheia was founded on a straightforward proposition: that the most valuable thing a person brings to a model is their judgement, and that most institutional settings are quietly organised to prevent them from using it. Approval chains, utilisation targets and the instinct to defer upward all have the same effect — they put distance between the person who can see the problem and the decision about what to do with it.
We are asking people to tell clients things those clients would sometimes rather not hear. That is only possible if the same freedom exists inside the firm, so we have built it in deliberately rather than hoping it emerges. What follows is less a list of benefits than a description of how a working week here actually goes.
We did not name the firm after a value we aspire to. We named it after the one we cannot do the work without.
MEASURED ON THE WORK
There are no utilisation targets here and no hours to fill. You are measured on whether the answer was right, whether it held up under challenge, and whether the client was better off for it.
It is a simpler basis to be judged on, and a more demanding one.
WORK YOU CAN PUT YOUR NAME TO
You take a piece of work from scoping through to the room where it is defended, and you meet the people it was built for.
It is a genuinely different experience of the same profession — you will be able to point at what you did and explain why you did it that way.
A SHORT DISTANCE TO A DECISION
Most things that need deciding can be decided the same afternoon, by the two or three people actually close to the work. There is no internal deck before the real deck.
Process earns its place here by making the work more accurate. Where it does not, we take it out.
CREATIVITY IS METHOD, NOT DECORATION
The creative act in this discipline sits in the framing: deciding what the real question is, what the model must refuse to assume, and where the standard approach stops working.
We would far rather you bring something unorthodox and argue for it than execute something conventional you do not believe in.
AUTHORITY FOLLOWS THE ARGUMENT
Seniority does not settle a technical disagreement; the evidence does. If you are right, you are right, and it does not matter who you happened to be right against.
The same freedom runs the other way. Saying “I am not sure” or “I had this wrong” costs nothing here, and it is usually the sentence that saves the engagement.
YOUR OWN VOICE, KEPT
We want you writing, speaking and holding positions under your own name — including positions the firm has not taken. Your professional reputation is yours to build, and it should travel with you.
If you eventually outgrow us, we would rather that happened in public.
— ROLES
JOIN OUR TEAM
The disciplines our engagements currently call for.
BUILD
Answerable for the model itself — its structure, its assumptions, its arithmetic.
Actuarial & Quantitative Modeller Builds the model. Owns the assumptions.
The work
- Build, extend and maintain reserving, capital, pricing and provisioning models to production standard — IFRS 17 measurement models, IFRS 9 expected credit loss models, SAM standard formula and internal model components, technical provisions and risk margin calculations.
- Own the assumption basis. Set it from data where data exists, derive it by analogy where it does not, and document clearly which of the two you did. Run and defend experience investigations across mortality, morbidity, lapse, expense, loss ratio and PD/LGD/EAD.
- Design model architecture that survives contact with reality: modular calculation layers, controlled inputs, versioned assumption files, and a clean separation between data, logic and presentation.
- Write model documentation as a deliverable, not an afterthought — methodology, assumption rationale, limitations, known weaknesses, and the circumstances in which the model should not be relied upon.
- Build and evidence controls: reconciliation to source and to prior period, movement and analysis of change, sensitivity and stress runs, and back-testing against emerging experience.
- Respond to independent challenge — from our validation teams, from external audit, and from the client’s own second line — in writing and in person.
- Work alongside client actuarial, finance and risk teams, frequently inside their environment and under their change-control regime.
What we look for
- A degree in actuarial science, statistics, mathematics or a comparably quantitative field, with credible progress through professional examinations — qualified, nearly qualified, or actively writing.
- Demonstrable model-building experience in insurance, banking or another regulated environment: reserving, capital, pricing or credit risk.
- Strong Excel and VBA, plus working fluency in at least one of Python, R, SAS or a vendor actuarial platform such as Prophet, ResQ, AXIS or Tyche.
- Working knowledge of at least one relevant framework — IFRS 17, IFRS 9, SAM, Basel, or a resources or infrastructure provisioning standard.
- The habit of showing your working: you can explain a number to a modeller, an auditor and a CFO without changing the substance for any of them.
- A preference for being wrong in private and corrected early over being wrong in a board pack.
Decision Analytics & Data Science Consultant Turns a dataset into a decision someone can defend.
The work
- Build forecasting, segmentation, propensity and pricing models for commercial and risk decisions — demand and volume forecasts, customer lifetime value, credit decisioning and collections prioritisation, churn and retention, claims frequency and severity.
- Do the unglamorous half properly: data profiling, lineage, missingness and bias assessment, feature construction, leakage checks, and an honest account of what the data can and cannot support.
- Select methods on the merits. A GLM the client’s team can maintain and interrogate frequently beats a boosted model nobody can explain; where complexity is warranted, justify it explicitly.
- Make outputs interrogable — SHAP or equivalent attribution, partial dependence, stability and drift monitoring, and confidence stated as a range rather than a point estimate dressed up as certainty.
- Translate output into a recommendation a CFO, CRO or exco can act on, including the recommendation not to act where the evidence is thin.
- Build for handover: reproducible pipelines, documented code, and a client team that can run, adjust and defend the model after we leave.
- Raise fairness and conduct risk early where a model materially affects individuals’ access to credit, cover or price — in the body of the report, not the appendix.
What we look for
- A quantitative degree — data science, statistics, actuarial science, econometrics, engineering or equivalent.
- Production experience in Python or R, comfort with SQL, and familiarity with version control and reproducible workflows.
- Practical grounding in supervised learning, time series and generalised linear models, with the judgement to know which the problem actually calls for.
- Experience with explainability tooling and model monitoring, or a clear appetite to work that way.
- The discipline to flag where a model’s confidence exceeds its evidence — including when the client would prefer you did not.
- Communication that survives contact with a non-technical audience without becoming untrue.
Project Finance & Infrastructure Modeller Builds the model the money is committed against.
The work
- Build and audit project finance models for infrastructure, energy and resources transactions — construction and operating phases, debt sizing and sculpting, cash waterfalls, reserve accounts, and full three-statement integration.
- Model the covenant architecture accurately: DSCR, LLCR and PLCR, lock-up and distribution tests, and the behaviour of the structure under breach and cure.
- Run sensitivity, scenario and stress analysis reflecting the risks actually present in the asset — construction delay and cost overrun, availability and offtake risk, tariff and indexation mechanics, currency and rate exposure, and macro assumptions specific to South Africa and the region.
- Support transactions end to end: base case agreement, lender due diligence and model audit responses, financial close, and post-close model maintenance and rebasing.
- Apply model integrity discipline — FAST or equivalent standards, no hard-coded constants, single-purpose rows, complete error-check panels, and an audit trail a lender’s technical adviser can follow without a walkthrough.
- Bring actuarial thinking into a corporate finance setting: distributional rather than single-point views of outcomes, explicit treatment of long-dated obligations, and honest discounting.
- Read concession agreements, PPAs, loan documentation and construction contracts closely enough to model what they actually say rather than what they are assumed to say.
What we look for
- A degree in actuarial science, finance, engineering, accounting or economics, and demonstrable project finance or infrastructure modelling experience.
- Advanced Excel modelling capability with structural discipline that holds up under third-party audit.
- Understanding of debt structures, funding competitions, DFI and commercial lender requirements, and the South African energy and infrastructure procurement environment.
- Exposure to model audit or lender’s technical adviser work is a strong advantage. A professional qualification — CFA, CA(SA), actuarial progress — is welcome but not required.
- Precision under transaction deadlines, including the willingness to say a number is not ready rather than let it go out unverified.
CHALLENGE
Answerable for finding what the build missed — before someone else does.
Model Validation Specialist The second pair of eyes that actually looks.
The work
- Perform independent validation of client models and, where mandated, of models built by other Aletheia teams — conceptual soundness, data quality, implementation correctness, assumption reasonableness, and fitness for the stated use.
- Design and execute the testing programme: replication of key results, sensitivity and stress testing, benchmarking against alternative methods, outcome analysis and back-testing, and behaviour at boundaries and extremes.
- Interrogate data lineage from source system to model input, including transformations, exclusions and manual adjustments. This is where most material findings actually live.
- Assess model risk in the round — governance, ownership, change control, version integrity, key-person dependency, and whether the documentation would let a competent stranger reconstruct the result.
- Validate machine learning and AI-assisted models specifically for explainability, stability, proxy discrimination and over-reliance, with reference to King governance expectations and POPIA obligations on automated decision-making.
- Write findings that are usable: severity-rated, evidenced, and paired with a remediation path — not a list of observations engineered to be unfalsifiable.
- Present and defend conclusions to model owners, audit committees, boards and regulators. Hold a finding under pressure where the evidence supports it, and withdraw it promptly and visibly where it does not.
What we look for
- Quantitative modelling literacy sufficient to rebuild what you are reviewing, not merely to comment on it.
- Experience in model validation, model risk, internal audit, second-line risk — or a build role you are now prepared to argue against.
- Familiarity with model risk management expectations: SR 11-7 principles, SAM and Basel model governance requirements, or a comparable internal framework.
- Strong technical writing. It is the most underrated skill in this discipline and the one we assess most closely.
- A temperament suited to independent challenge — treating “I cannot follow why this works” as a finding rather than a personal shortfall, and separating the model from the modeller.
- Enough diplomatic range to deliver an unwelcome conclusion without turning it into a fight.
GROUND
Answerable for the model’s contact with regulation, and with the real asset.
Regulatory & Capital Reporting Specialist Knows what the regulator will ask before they ask it.
The work
- Lead or support regulatory capital and reporting workstreams — SAM own funds and SCR calculation and reporting, ORSA production and review, Basel capital and disclosure requirements, and IFRS 17 and IFRS 9 disclosure preparation.
- Run capital readiness assessments: quantify the position under current and stressed conditions, identify the calculation and control gaps a supervisory review would surface, and sequence remediation against real reporting deadlines.
- Prepare and quality-assure regulatory submissions and supporting technical memoranda, and manage the query cycle with the Prudential Authority, the FSCA or the relevant supervisor.
- Translate regulatory developments into concrete client implications — not a summary of the circular, but what it changes in the calculation, the governance and the timetable.
- Sit between actuarial, risk and finance and reconcile their numbers to each other, which is frequently where the real work is.
- Draft board and audit-committee material that states the position plainly, particularly where the position is uncomfortable.
- Support technical provisions and capital assurance work alongside external auditors and reviewing actuaries.
What we look for
- An actuarial, risk management or financial reporting background with direct exposure to a regulatory capital regime — SAM, Solvency II, Basel or equivalent.
- Hands-on experience producing or reviewing regulatory returns, ORSAs, or IFRS 17 and IFRS 9 disclosures under a real reporting deadline.
- Fluency in the South African regulatory landscape: the Insurance Act and prudential standards, FSCA conduct expectations, the King codes, and POPIA where models touch personal information.
- Comfort operating close to a board, audit committee or exco audience.
- Precision with detail, alongside the judgement to know which details matter to a supervisor and which do not.
Sector Subject-Matter Advisor — Mining, Resources & Industry Understands the asset, not just the spreadsheet.
The work
- Provide domain grounding on engagements in mining, resources and heavy industry — closure and rehabilitation provisioning, engineering labour and skills modelling, production and throughput forecasting, occupational health liability estimation, and capital project appraisal.
- Sense-check model outputs against physical and operational reality: mine plans and life-of-mine schedules, shaft and plant constraints, grade and recovery profiles, labour complements and shift patterns, maintenance and availability, and rehabilitation scopes as actually costed.
- Interrogate assumptions that look plausible in a spreadsheet and are impossible underground. This is the single highest-value contribution the role makes.
- Work as the translator between technical model and technical asset, alongside actuarial, validation and analytics colleagues who need operational context they do not have.
- Advise on the applicable standards and regulatory environment — MPRDA financial provision requirements, NEMA closure obligations, mine health and safety regulation, and industry reporting standards.
- Contribute at scoping and risk-identification stage, where sector knowledge changes the shape of the work rather than merely reviewing it at the end.
- Support engagement with operational management, engineering and HR functions, who are frequently the source of the data the model depends on.
What we look for
- Substantive sector experience — mining operations, mining engineering, metallurgy, environmental and closure planning, industrial workforce planning, or resources corporate finance.
- Enough numerical fluency to work productively alongside modellers and to challenge a number on its own terms.
- Credibility with operational management as well as with head office.
- Professional registration — ECSA, SACNASP, SAIMM membership or similar — is an advantage.
- Openness to a project-based or fractional arrangement alongside existing commitments.
CARRY
Answerable for delivery, for the client relationship, and for the next generation of the discipline.
Engagement Lead Keeps the work honest and on time. Both.
The work
- Own delivery of an engagement end to end — scope definition, workplan, resourcing, budget, quality of output, and the client relationship at CFO, CRO, head of actuarial or head of risk level.
- Compose and coordinate the team across modelling, validation, sector advisory and analytics, and make the trade-off calls when their priorities conflict.
- Hold the line on scope and standard simultaneously. Under deadline pressure something gives; this role decides what, and it is not the rigour.
- Run the review chain. Nothing leaves without a named reviewer, a documented review, and a resolved comment log.
- Manage the difficult conversations: a finding the client did not want, a timeline that no longer holds, a scope that has quietly doubled, an assumption management is attached to that the evidence does not support.
- Contribute to proposals, scoping conversations, published technical writing, and the firm’s capability and assurance transfer work — leaving client teams more capable than we found them.
- Develop the analysts on the engagement deliberately: review their work in detail, explain the corrections, and give them exposure they are only just ready for.
What we look for
- A consulting, actuarial function or risk delivery background in financial services, insurance or resources, with genuine ownership of engagements or workstreams.
- Technical depth sufficient to review the work, not merely to project-manage it. This is not an oversight role.
- Evidence of holding a standard under commercial pressure — and of having lost that argument at least once and understood why.
- Strong written and verbal communication with senior stakeholders, and the ability to be direct without being combative.
- An interest in building a firm rather than filling a seat in one.
Graduate Analyst Where most of this team started.
The work
- Enter the discipline through a structured development and skills-transfer programme, rotating across modelling, validation and analytics work under close senior supervision.
- Do real client work early: data preparation and validation, experience investigations, model testing and replication, sensitivity runs, documentation and review support. Not simulations, and not slide formatting.
- Learn the firm’s technical standards explicitly — how a model is structured, how an assumption is justified, how a finding is written, and how a number is checked before it leaves the building.
- Build toward an actuarial or professional qualification with study support, protected study time, and a mentor who is accountable for your progress.
- Contribute to internal knowledge assets: technical notes, model documentation, and research supporting the firm’s published work.
- Be reviewed candidly and often. Corrections here are detailed and unsentimental, and they are the point of the programme.
What we look for
- A completed or near-complete degree in actuarial science, statistics, mathematics, data science, economics, engineering or a comparable quantitative field.
- A strong academic record, and Excel capability beyond the introductory. Coding exposure — Python, R, SQL, VBA — is an advantage; none is disqualifying.
- Evidence that you finish things: a completed project, thesis, exam sequence, or something built outside a curriculum.
- Clear written English. You will be writing for people who read carefully.
- The appetite to be taught rigorously, and to be told plainly when something is wrong.
— OUR VALUES
WHAT WE HOLD IN COMMON.
These are not values in the poster sense. They are the dispositions the work genuinely requires, and the ones we look for in the people who join us — the kind a person tends to carry into the rest of their life rather than put on at the door.
01
Truth
A preference for what is real over what is comfortable, including when the truth is about our own work. People here would rather know than be reassured, and say the accurate thing when an easier one is available.
02
Excellence
Care for the work itself rather than the appearance of it. The standard that counts is the one you keep in the parts nobody will ever inspect, and for most people here it is a habit rather than an effort.
03
Service
Work is always done for somebody. We look for people who keep that person in view — the client, and past them whoever the decision finally lands on, who is rarely in the room and rarely able to argue back.
04
Courage
Saying the difficult thing early and plainly, and saying it with enough care that it can actually be heard. The same courage is needed to say “I had this wrong,” which is usually the harder of the two.
05
Accountability
Owning what you put your name to, from the first assumption through to the final sign-off. It also means doing what you said you would do in small things — the reliability colleagues build their plans around.
— The process
HOW AN APPLICATION MOVES.
You apply
Submit your CV against the discipline closest to your background, with a short note on the work you want to be doing. We read the note.
We review
Every application is read by someone who does the work, not filtered on keywords.
Technical conversation
A discussion of a real problem in your discipline. We are interested in how you reason, and what you do when you are unsure.
Fit and terms
A conversation with the principal about the firm, the standard, what you want from your career, and the terms on which you would join us.
— APPLY
SEND US YOUR CV.
Tell us what you want to be accountable for.
Applications are reviewed on a rolling basis across insurance, banking, mining and resources, project finance and infrastructure, and retail.
If your background sits across two disciplines, choose the one you would want to be measured on, and say so in your note.
ALETHEIA PARTNERS
Analytics. Modelling. Truth.
A specialist analytics and modelling consultancy, grounded in actuarial science. Based in Johannesburg, working with corporates, project sponsors and investors across Africa.
© 2026 Aletheia Partners (Pty) Ltd · All Rights Reserved
SPACE