Suraj Thapa

St. Louis, Missouri

208-310-4215

Interests

Machine Learning , MLOps , LLMs/RAGs/Fine tuning LLMs/NLP/ , Computer Vision

Skills

Languages: Python, Bash, Scala, R

LLM Experiences: Retrieval-augmented generation (RAG), Fine tuning, OpenAI, Cohere, Gemini, Agent Builder, Weaviate, GPT APIs, Langchain, Streamlit, LLamaindex, Huggingface

Database: PostgreSQL (pgvector), Weaviate, MySQL, CrateDB, MongoDB

AWS: AWS Bedrock, AWS Sagemaker

GCP: Document AI, Vertex AI, Agent Builder

Git: GitHub, GitLab, CI/CD pipelines

Infrastructure as Code: Terraform, Cloudformation

Observability Tools: Sentry, New Relic, Papertrail

Containerization: Docker, Kubernetes, ECS, EKS

Other Experiences: scikit-learn, pyTorch, NumPy, Pandas, Linux, Ansible, Docker, Web, Kafka, RabbitMQ, AirFlow, Databricks, Web scraping, Jira

Education

University of Denver, Denver, CO

M.A., Global Economic Affairs

2018 - 2021

University of Idaho, Moscow, ID

B.A., Bachelor in Economics; Minor in Math and Statistics

2015 - 2018

Work

Machine Learning Engineer, The Texas Tribune

www.texastribune.org

Developed AI-powered chatbots adhering to journalistic standards, ensuring accuracy and minimizing hallucinations.
Implemented hybrid search systems using PostgreSQL for internal applications, integrating semantic, fuzzy, and full-text search for Retrieval-Augmented Generation (RAG).
Designed a cross-encoder re-ranker for RAG applications, significantly improving document retrieval accuracy and relevance.
Fine-tuned transformer models (e.g., BERT for Named Entity Recognition) from Hugging Face for internal applications.
Designed and managed scalable ML infrastructure for model training, testing, and deployment in cloud-native environments.
Created automated CI/CD pipelines for ML workflows, including data ingestion, feature engineering, model training, validation, and deployment on AWS and GCP.
Established observability frameworks for end-to-end monitoring of model performance, data quality, and system reliability.
Implemented infrastructure-as-code (IaC) solutions using Terraform and CloudFormation to maintain reproducible, version-controlled production environments.

2022 - Present

Data Engineer II, Lightcast

lightcast.io

Built Machine Learning Models to classify web pages to enhance web scraping capabilities
Implemented and maintained GitLab CI/CD workflows and Code Deploy (AWS) using container services in GitLab
Performed data migration from Crate database to AWS RDS, PostgreSQL
Utilized docker to build and deploy applications

2021 - 2022