Mercor resume generator
Mercor screens your resume with AI before the interview, and the interview then follows what your resume claims. This generator turns your real experience into a one-page, single-column resume that puts your narrow domain expertise first, then downloads as a PDF.
AI Work Expert is independent and not affiliated with Mercor.
Paste anything about you
Notes, an old resume, a LinkedIn dump - anything. The AI turns it into a clean, ATS-plain one-pager.
Powered by AI. Nothing is stored - the text you paste is sent only to generate your resume. Connect your own GPT/Claude key from the admin panel.
Summary
Machine learning engineer with 7+ years building and evaluating production ML systems, now focused on frontier-model training data and evaluation. Experienced in designing rubrics, auditing model reasoning, and leading annotation quality programs at scale. Known for turning ambiguous evaluation criteria into measurable, reproducible standards that research teams can act on.
Core Competencies
Model Evaluation: Rubric design, pairwise preference ranking, RLHF feedback loops, reasoning-trace audits, red-teaming, inter-rater reliability
Machine Learning: PyTorch, Hugging Face Transformers, fine-tuning, LoRA / PEFT, retrieval-augmented generation, model benchmarking
Data & Analysis: Python, NumPy, pandas, SQL, statistical testing, sampling design, data pipelines, Airflow
Engineering: Distributed systems, REST / gRPC APIs, Docker, Kubernetes, AWS, CI/CD, code review at scale
Collaboration: Cross-functional research partnership, annotation team leadership, technical writing, async documentation
Professional Experience
Lead the evaluation workstream for a frontier language model, owning how reasoning quality is measured across coding, math and domain-expert tasks.
- Designed the evaluation rubric used by 120+ expert annotators, raising inter-rater agreement from 0.61 to 0.87 kappa.
- Built an automated triage pipeline that surfaces low-confidence model outputs for human review, cutting review volume 45% with no loss in coverage.
- Audited agent execution traces across multi-file Python and Go repositories to flag reward hacking and unsafe tool calls.
- Partnered with research to convert qualitative reviewer feedback into training signal, contributing to a 12-point gain on internal reasoning benchmarks.
- Mentored six junior engineers and established the team's code review and documentation standards.
Built and shipped production ML services for enterprise customers across forecasting, classification and document understanding.
- Shipped eight production models serving 40M+ monthly inference requests at p99 latency under 120ms.
- Led a data quality initiative that reduced label noise by 30%, improving downstream model F1 by 9 points.
- Designed and ran A/B tests for model rollouts, establishing the company's standard evaluation playbook.
- Migrated the training stack to containerised pipelines, reducing experiment turnaround from days to hours.
- Delivered forecasting and segmentation models for retail and logistics clients.
- Automated recurring reporting workflows, saving an estimated 20 analyst hours per week.
Education
Certifications & Continuous Learning
- Responsible AI: Evaluation & Red-Teaming Practices
- Deep Learning Specialization - Neural Networks & Sequence Models
- AWS Certified Machine Learning - Specialty
- Continuous Learning: LLM Alignment & Preference Optimization
How this works
Paste your notes, an old resume or a LinkedIn export. An AI model structures it into our one-page layout with a Mercor emphasis on the headline and the order of skills. It uses only the facts you paste. Check every line before you submit it: a submitted application cannot be edited on Mercor, and the interview will ask about what the resume says.