Vikram Krishnan
AI Research EngineerEmail:vikram@vk-ai.dev
Phone:+1 415 555 0173
Location:San Francisco, CA
LinkedIn:linkedin.com/in/vkrishnan
GitHub:github.com/vkrishnan
Professional Summary
AI engineer with 6 yrs training and deploying LLMs, retrieval systems, and multi-modal models. Published at NeurIPS + ICLR. Ex-startup CTO. Known for clear communication, ownership, and shipping measurable outcomes. Combines strong data fundamentals with cross-functional collaboration and a bias for execution.Experience
Aug 2023Present
AI Research Engineer
OpenAI · San Francisco, CA
• Contributed to GPT-4.5 post-training data pipeline (10k+ eval sets)• Reduced inference cost 34% via speculative decoding kernel optimization
• Co-authored 2 papers on constitutional AI (in review, ICLR 2026)
• Built models/pipelines at OpenAI impacting $9M+ decisions annually
• Reduced reporting latency from days to hours for 10 stakeholder teams
Oct 2021Jul 2023
Founding Engineer
Perplexity AI · San Francisco, CA
• Employee #4; built core retrieval pipeline serving 100M+ queries/month• Shipped Pages feature contributing to $50M ARR run rate
• Documented analysis playbooks used across the org
• Built models/pipelines at Perplexity AI impacting $4M+ decisions annually
• Reduced reporting latency from days to hours for 11 stakeholder teams
Skills
Python, PyTorch, JAX, CUDA, Transformers, RAG, LangChain, LlamaIndex, vLLM, Ray, Distributed Training, RLHF, Evaluation, FastAPIProjects
nanoRAG
Python, PyTorch
github.com/vkrishnan/nanorag
spec-decode-cuda
CUDA, C++
github.com/vkrishnan/spec-decode-cuda
Experimentation Platform (OpenAI)
Python, SQL, dbt, Spark, Tableau, Airflow
• Documented learnings reused by peer teams
• Recognized in team/org review for execution quality
Experimentation Platform (OpenAI)
Python, SQL, dbt, Spark, Tableau, Airflow
• Documented learnings reused by peer teams
• Recognized in team/org review for execution quality
Education
M.S, Computer Science
Stanford University
Stanford, CA
Focus: NLP & Deep Learning · GPA 3.95/4.0 · Stanford AI Fellowship