Nikzad Khani

Software Engineer.

Building scalable AI-driven applications and robust data pipelines with Go and Python.

View ExperienceContact Me

Expertise

Languages

Frontend & Frameworks

Cloud & Infrastructure

Data & AI

Python

4.2+ Years

Applied Context

Senior Software Engineer
Present
Verily
Software Engineer III
May 2025
Verily
Software Engineer I
June 2023
PathAI
Researcher and Coordinator
August 2020
AI4ALL
An Exploration of Deep Reinforcement Learning Methods in Hungry Geese
2021
Publication • ArXiv
Cultural and Geographical Influences on Image Translatability of Words across Languages
2021
Publication • NAACL

Professional History

Senior Software Engineer

May 2025 — Present
  • Engineered and launched two distinct LLM-powered features for the Verily Me mobile application, enabling users to query personal healthcare data via natural language.
  • Architected and deployed scalable, agentic AI models to production on Google Kubernetes Engine (GKE), leveraging GCP Vertex AI and LangGraph.
PythonGoTypescriptReactGCPTerraformKubernetesLLMsLangGraphVertex AIDockerSingle SPAGithub CI/CD

Software Engineer III

July 2023 — May 2025
  • Built and orchestrated robust data pipelines with Airflow, processing hundreds of thousands of FHIR records daily into BigQuery.
  • Developed a full-stack microservice using gRPC and React to provide embedded Looker dashboards.
  • Automated the deployment of all GCP-based infrastructure using Terraform and configured GitHub workflows for CI/CD.
PythonGoTypescriptReactGCPTerraformKubernetesAirflowBigQuerySQLgRPCLookerDockerLangGraphSingle SPAGithub CI/CD

Education

Boston University

Bachelor of Arts in Computer Science

2017 — 2021

Publications

An Exploration of Deep Reinforcement Learning Methods in Hungry Geese

September 2021
ArXiv
  • Explored the effectiveness of using Deep Q-Networks with various architectures in a stochastic multiplayer snake environment.
  • Showed how optimizations to feature engineering could make DQNs more effective, although they are outperformed by Proximal Policy Optimization-based methods.

Cultural and Geographical Influences on Image Translatability of Words across Languages

June 2021
NAACL
  • Developed two metrics to show when images could be useful for machine translation.
  • Concluded that when speakers of two languages share culture, images are more likely to be useful for machine translation compared to shared Geography or language families.