Senior ML Engineer - Malware Detection (remote work)

Brak informacji o wynagrodzeniu

SeniorZdalnie·WarszawaFull-time

#466183·Dodano 2 dni temu·1

Źródło: Cloudlinux

Aplikuj teraz

Tech Stack / Keywords

Python PyTorchHugging FaceSQLClickHouseONNXvLLMPHP JavaScriptLoRA

Podsumowanie oferty

Build and ship malware detection for Imunify360 using production web-server data. The role covers ML models, LLM-assisted triage, and automatically generated detection signatures, with model performance measured in production.

Wymagania

5+ years building ML systems that ran in production, including 2+ years in an adversarial domain such as security, abuse, fraud, or spam.
Personally trained and fine-tuned transformer models, including data preparation, code tokenization, LoRA/QLoRA or full fine-tuning, distillation, and quantization; experience limited to calling hosted APIs is not sufficient.
Strong classic ML knowledge, including gradient boosting, linear models on sparse features, calibration, class imbalance, and judging when these approaches outperform an LLM.
Evaluation experience with leakage-free splits by time and family, precision at a fixed false-positive budget, regression sets, and accounting for differences between offline evaluation and production.
Experience measuring inference throughput and latency and fitting a model to CPU or memory budgets using ONNX, llama.cpp, vLLM, or int8/int4 quantization.
Strong Python, PyTorch, and Hugging Face stack skills; SQL on large analytical tables such as ClickHouse or similar.
Familiarity with PHP or JavaScript, or willingness to learn one while working with malware analysts.
Daily use of AI coding agents and ability to demonstrate how you use them.
English at upper-intermediate (B2) level or higher, with particularly strong written communication for asynchronous work.

Nice to have:

Experience with web malware, webshells, phishing kits, or WordPress security.
Experience with models over code, including AST or token features, code LLMs, or graph models on program structure.
Malware clustering and similarity analysis experience.
Multi-GPU training on your own bare metal; MLflow or Weights & Biases; dataset versioning.
Go or Rust for on-server inference.
Published work or talks on ML for security.

Obowiązki

Detection and model development:

Prioritize files on servers for closer inspection within CPU and memory budgets.
Classify malicious and benign code across the server fleet while managing cost per file.
Generate detection signatures automatically from new campaigns.
Detect malware that appears malicious only in context, including phishing pages and script injections.
Route files to the appropriate detection engine.
Improve cleanup precision so sites remain clean.
Prove that each model improves detection before release.

First month:

Train an initial model on production data and measure it against current detection engines.
Establish an evaluation harness with splits by time and campaign, and a regression set based on past false positives and false negatives.

By month 3:

Put several models into production, including models in shadow mode and models producing detections, gated on the past-error regression set.
Establish a training setup that keeps the GPUs busy.
Write an ADR proposal for the detection cascade, including cost per file and hardware needs.

By month 12:

Make models a main source of new detections, reporting their share of new detections and the false-negative-rate trend.
Retrain on fresh analyst and false-negative labels on a fixed cadence.
Integrate generated signatures into the analyst workflow.
Enable other engineers to run two or three parallel training directions using your tooling.

Benefity

Fully remote work with flexible working hours, allowing you to work from any location worldwide.
24 paid vacation days per year, 10 days of national holidays, and unlimited sick leave.
Compensation for private medical insurance.
Co-working and gym/sports reimbursement.
Budget for education.
Opportunity to receive a reward for the most innovative idea that the company can patent.

Elastyczne godziny

Płatne święta

Ubezpieczenie

Dofinansowanie szkoleń

CloudLinux

6 aktywnych ofert

Zobacz wszystkie oferty

Aplikuj teraz