Hello!

I'm Rishab Mallick, an AI product engineer

Get in touch rishabmallick6@gmail.com

Background

I'm currently working at metaintro, building job-matching, conversational search, and evaluation systems.

Previously, I worked at Unify (YC W23) on LLM benchmarking, provider routing, and open-source ML systems, and at Small Joys on health-focused NLP and clinical data tooling.

Experience
AI Product Engineer @ metaintro
Apr 2024 - Present
  • Built job-matching systems that rank roles against resumes, goals, and user constraints.
  • Built conversational AI backends for context-preserving job search, streamed job cards, and resume artifacts.
  • Built production quality loops for conversation audits, retrieval checks, model comparisons, cost and latency tracking, dashboards, failure attribution, and regression cases.
Machine Learning Engineer @ Unify (YC W23)
May 2023 - Apr 2024
  • Designed LLM benchmarking methodology and CI/CD runtime benchmarks on Google Cloud.
  • Built provider routing across Perplexity, Replicate, OpenAI, Anyscale, Vertex AI, and other model endpoints.
  • Integrated Ivy into BentoML, Seldon-Core, Seldon-MLServer, and MMPretrain while contributing to cross-framework transpilation work.
Founding AI Engineer @ Small Joys
Nov 2021 - Aug 2023
  • Built bias-robust sentiment models, AWS inference APIs, and behavioral-health AI prototypes.
  • Built journal abstraction and theme-extraction workflows for therapist-facing clinical context.
  • Built ScreenerOMR plus labeling and GPT-3 augmentation workflows for clinical data.
Student Software Developer @ Google Summer of Code
Jun 2021 - Aug 2021
  • Built a BERT, regex, and dictionary-based biomedical NLP extraction framework for WormBase C. elegans literature.
  • Helped reduce manual biocuration from hours to minutes and recover variant data missed in manual curation.
Selected early AI & data internships
Aug 2020 - Jun 2021
  • Built masked-face recognition, wearable object detection, e-proctoring, CAD comparison, and geo-data cleanup prototypes across CDAC, Myraa Technologies, and Tata Steel.
Résumé (PDF)
Others
Dr. APJ Abdul Kalam IGNITE Award

Received IGNITE 2015 from the National Innovation Foundation - India for "Pay as you weigh." Indian Patent No. 531470 was granted in 2024.

National Innovation Foundation ·

Skills
Relevant Coursework
Technical
  • Python
  • Machine learning
  • Natural language processing
  • LLM evaluation
  • Applied AI systems