AI/ML Engineer

Jwalix is an AI-powered applicant tracking platform for in-house recruiting teams and staffing agencies who want one workspace instead of stitched-together tools — jobs, candidates, interviews, placements, campaigns, and reporting, all in one place, with an AI Recruiting Copilot built into the workflow rather than bolted on. Read more about Jwalix →

Jwalix's AI features are built to act inside the existing workflow rather than as a separate chat tab: an AI Recruiter that works through the same modules and permissions a user has, semantic matching that runs locally alongside keyword search, and asynchronous AI interviews scored against a structured question plan. You'd work across all three.

This isn't a research role sitting apart from the product — every model or pipeline you build has to run inside a live, multi-tenant SaaS product with real latency and cost constraints, and ship to actual recruiters, not stay in a notebook.

You'd report to the Head of AI and work closely with backend engineers on integration and with Customer Success on what customers actually need.

• Improve semantic candidate matching — Tune and evaluate the embedding-based ranking that runs alongside keyword search, with a clear fallback path when it's unavailable.
• Extend the AI Recruiter — Build and refine the tool-calling layer that lets the assistant search, summarize, and draft outreach through real modules — under the requesting user's own permissions.
• Work on AI interview scoring — Improve how an async interview transcript gets evaluated against the interviewer's own question plan and rubric.
• Own evaluation, not just prompts — Build the offline evaluation sets and metrics that tell you whether a change actually helped, before it ships to every tenant.
• Manage cost and latency — Make deliberate tradeoffs between hosted LLM calls and smaller local models, so AI features stay fast and affordable at scale.
• Document what the AI does and doesn't do — Write clear internal documentation of model behavior and limitations, since Customer Success and Sales rely on it to set accurate expectations.

• 3+ years applying ML/NLP in a production product, not only research
• Hands-on experience with embeddings and semantic search or ranking
• Comfortable working with hosted LLM APIs and structured tool-calling
• Able to reason about latency, cost, and failure modes for AI features in a live product, not just accuracy
• Comfortable with Python and standard ML tooling for evaluation and experimentation
• Able to explain a model's behavior and limitations clearly to non-ML teammates

• Experience running smaller models on CPU for cost- or privacy-sensitive steps
• Background in recruiting, search, or recommendation systems
• Experience with voice/speech pipelines (relevant to our AI voice screening feature)

Apply Now

• 1. Application review — We read every application. Expect to hear back within 2–3 business days either way.
• 2. Intro call — A 30-minute conversation with the hiring manager about the role and your background.
• 3. Technical round — a practical exercise using a dataset similar to what you'd actually work with here
• 4. Team interview — Meet 2–3 future teammates and go deeper on how you'd work together.
• 5. Offer — If it's a fit on both sides, we move quickly rather than dragging out a final decision.

Jwalix is an equal opportunity employer. We welcome applicants from all backgrounds, and hiring decisions are based on merit and fit for the role — not race, gender, religion, age, disability, or any other protected status.