Chris Samanant
Twin Cities (MN, USA) · chris.samanant.jobs@gmail.com
Data ScienceApplied MLRetrievalEvaluationVespaGCP · GKE
Why I built this: Rather than a static résumé, this is a living portfolio that actually runs the systems I design—data ingestion, enrichment, ranking, evaluation, and monitoring. It’s the strongest demonstration of data-science craftsmanship a hiring manager can see.
Career Statement
I build full-stack data systems that learn, adapt, and deliver measurable value. My experience bridges data engineering, applied machine learning, and collaborative feedback loops, ensuring that models stay grounded in real-world performance and human understanding. With a foundation in mathematics, statistics, and software engineering, I take end-to-end ownership—from raw data to deployed intelligence—to produce results that are explainable, reliable, and aligned with human progress.
Education
University of Minnesota — Twin Cities
- MS in Business Analytics — Carlson School of Management (2020)
- BS in Economics (2019)
Professional Experience
Full-Stack Data Scientist — Independent (Mar 2025 – Present)
Vector search R&D · Consensus eval · GKE/Vespa
Built daily ingestion on GKE with OpenAI enrichment and Vespa vector search. Shipped a consensus-driven evaluation stack (P@K, MAP, nDCG) and multiple live demos (intro search, single-search, fine-tune). Deployed the unified platform at HeyBye.ai.
Senior Data Scientist — Optum (Jan 2022 – Mar 2025)
Benefits Search · NLP retrieval · Cloud deployments
Designed a Vespa application with multi-vector NLP search; evaluated models and rankers with automated Python tooling and dashboards; collaborated on an OpenAI-powered QA assistant using retrieved context.
Data Engineer — Technology Development Associate (Jun 2021 – Jan 2022)
SQL/Snowflake migrations · Procedure optimization
Validated large data migrations; optimized stored procedures; built operational monitors to catch data quality issues.
Analytical Student Consultant — Carlson Analytics Lab (Jun 2019 – May 2021)
Client analytics · Causal analysis · Forecasting
Delivered descriptive & causal analytics for industry partners; presented recommendations and led final client briefings.
Technical Skill Map
| Domain | Tools / Frameworks |
|---|---|
| Languages | Python, SQL, YAML, Bash |
| ML & AI | OpenAI (GPT-4o family), Hugging Face, scikit-learn, PyTorch embeddings |
| Search & Retrieval | Vespa.ai, BM25/ANN rank profiles, evaluation metrics |
| Data Pipelines | FastAPI, GCP (GCS, Pub/Sub), Pandas; Airflow (planned) |
| Infra / DevOps | Kubernetes (GKE), GitHub Actions, Cloud Build, Docker, Grafana + Prometheus |
| Databases | PostgreSQL (CloudSQL), SQLite for demos; Redis (planned) |
| Visualization & UX | Chart.js / D3, Tailwind, Next.js |
Projects Showcased on This Site
Arxiv → Vector Docs
NLP enrichment & embeddings
GPT-4o pipelines for summaries, keywords, labels, and vectors.
Vespa Vector DB Namespace
Multi-node GKE cluster
Custom rank profiles, ANN + keyword hybrid, YQL playground.
Auto-Labeler + Consensus
Human-in-the-loop labels
GPT priors, Bayesian posterior with rater weights, priority queues.
Endorsement
I worked with Chris at Optum, where he was an awesome Al/ML engineer. Chris is a super talented and emerging technology leader, who has made game-changing contributions to our company's Al/ML initiatives.
Chris was the lead engineer for the Proof of Concept project using NLP models as a cornerstone in a Benefits Search Engine. This project aimed to help our customers find their plan's benefit documents easily and accurately, based on any search phrase. Chris came up with and built the NLP semantic search algorithms, using Vespa as the chosen vector database and application platform. He also used OpenAl's LLMs to provide Benefit Answering on top of Benefit Search, using Few-Shot, Role-Based Prompt Engineering. This feature let our customers ask questions about their plan and get a chatbot's conversational answer, instead of reading multiple coverage documents from start to finish.
Chris's work on this project was amazing and impressive. He showed his skills in NLP, ML, and cloud computing, as well as his creativity and problem-solving skills. He also kept the solution on our company servers with K8s, and made a neat user interface around the searcher using Streamlit. He showed his work to various stakeholders and got positive feedback and recognition. His statistical analytics showed that his PoC Search algorithm usually connected patients with more than twice as many relevant benefits as the enterprise's previous solution, repeatedly seen in our standardized search evaluation metrics like Precision@K, Recall@K, and MAP.
With demonstrated potential, Chris was also part of scaling up and providing a production grade API dedicated to Benefit Search with NLP for almost 100 million patients. He used tools like Helm, Terraform, and RunIAC to set up and network our containerized application on Azure Kubernetes, with the needed security, authorization credentials, and CI/CD capabilities. He also worked well with other engineers and teams, and talked effectively with managers and clients. He was always proactive, reliable, and professional in his work.
Chris is awesome at using Al/ML technologies to make better products and services. He is really passionate and talented in his field, and I'm sure he will keep rocking in his future projects. Hiring Chris is a no-brainer, as he is a great candidate for any job that he wants to do.
Chris was the lead engineer for the Proof of Concept project using NLP models as a cornerstone in a Benefits Search Engine. This project aimed to help our customers find their plan's benefit documents easily and accurately, based on any search phrase. Chris came up with and built the NLP semantic search algorithms, using Vespa as the chosen vector database and application platform. He also used OpenAl's LLMs to provide Benefit Answering on top of Benefit Search, using Few-Shot, Role-Based Prompt Engineering. This feature let our customers ask questions about their plan and get a chatbot's conversational answer, instead of reading multiple coverage documents from start to finish.
Chris's work on this project was amazing and impressive. He showed his skills in NLP, ML, and cloud computing, as well as his creativity and problem-solving skills. He also kept the solution on our company servers with K8s, and made a neat user interface around the searcher using Streamlit. He showed his work to various stakeholders and got positive feedback and recognition. His statistical analytics showed that his PoC Search algorithm usually connected patients with more than twice as many relevant benefits as the enterprise's previous solution, repeatedly seen in our standardized search evaluation metrics like Precision@K, Recall@K, and MAP.
With demonstrated potential, Chris was also part of scaling up and providing a production grade API dedicated to Benefit Search with NLP for almost 100 million patients. He used tools like Helm, Terraform, and RunIAC to set up and network our containerized application on Azure Kubernetes, with the needed security, authorization credentials, and CI/CD capabilities. He also worked well with other engineers and teams, and talked effectively with managers and clients. He was always proactive, reliable, and professional in his work.
Chris is awesome at using Al/ML technologies to make better products and services. He is really passionate and talented in his field, and I'm sure he will keep rocking in his future projects. Hiring Chris is a no-brainer, as he is a great candidate for any job that he wants to do.
— Andrei Filimonov, Manager, Optum
About Me
I like lifting, running, and getting so sore I can barely move. I like eating healthy food so fast I could pop. I like music that envelops you fully within the present moment. I like listening to or reading philosophy, science fiction, and social commentary — thinking critically about what is and what could be. I like sports — any sport, any day — clock and scoreboard turned off. I like digital curation — collecting ideas, videos, fragments — the modern way to build a museum of curiosity... clips, quotes, fragments, highlights... catching sparks before they vanish. I like Disney, Jedi, and Superheros. I like games of strategy and timing — Rocket League, Fortnite, Chess... and "even though I do not know you, and even though I may never meet you, laugh with you, cry with you, or kiss you. I love you. With all my heart, I love you.”
Contact / Availability
- Email: chris.samanant.jobs@gmail.com
- Location: Remote / Hybrid (Minnesota or U.S.-based)
- Availability: Open to Full Stack Data Scientist Roles
- Links: LinkedIn · GitHub
0:00 / 0:00