Skip to content

Experience

AI engineer with ~3 years shipping production AI systems, across healthcare ( HEOR / SLR extraction at Bridge Medical), finance (real-time + long-form equity research at KalqylAI / Predli), and enterprise SaaS (RAG and document ingestion at Predli Studio). Most recently at Bridge Medical (Red Nucleus); open to AI engineering roles and collaborations.

Roles

Dec 2025 → Jun 2026

AI Engineer

Bridge Medical Consulting (a Red Nucleus company)

  • Led the AI technical and deployment strategy for HEOR Systematic Literature Review extraction across two parallel tracks: hardening the existing notebook workflow into a tool the whole team could use, and shaping a structured, audit-first extraction approach.
  • Track 1 — productionizing the existing extraction workflow: automated the manual steps, containerized the Jupyter-notebook pipeline, and exposed it through an interactive UI for non-technical reviewers, cutting turnaround roughly in half.
  • Track 2 — designing a structured extraction approach for ~100 interdependent fields per study, with smart ingestion, staged validation, agentic escalation on low-confidence fields, and citation-level provenance.
  • Drove the technical choices — models, agentic framework, data model, and tracing — with a bias toward auditable systems domain experts can trust.
2023 → 2025

AI & ML Consultant

Predli Consulting AB

  • Owned end-to-end build of KalqylAI for Spotlight Group across all three layers: real-time financial data infrastructure (160+ Swedish-listed companies on Azure), the Twitter agent (@KalqylAI) reacting to press releases in under 30 seconds, and the analytical bot publishing long-form AI equity research reports. 8,000+ reports and 4.6M+ annual views in production.
  • Designed the data-layer-as-product architecture that decoupled WebSocket ingestion, multi-format document parsing, and a Naive Bayes relevancy/PR classifier from the bot consumers, so both bots ran on the same enriched-event contract.
  • Built entity-grounding validation into the Twitter bot: ticker, company, and subsidiary references in candidate tweets reconciled back against the structured data-layer payload before publish.
  • Built the Predli Studio enterprise document ingestion pipeline on Dagster: asset-based orchestration over a task-based scheduler, with content-hash idempotency at the raw-document boundary so connector retries don't poison the vector store.
  • Built the Enterprise RAG platform behind Studio with file-type-aware chunking (PDFs, decks, spreadsheets), TOC-based hierarchical context, and persona-aware retrieval. 40% relevance improvement over uniform-chunking baseline.
2023 (Apr → Jun)

Data Science Intern

Predli Consulting AB

  • Developed end-to-end NLP pipeline for survey response analysis.
  • Implemented clustering algorithms using Huggingface transformers, UMAP, and HDBSCAN.
2019 → 2022

Business Operations Associate

ZS Associates

  • Designed incentive compensation frameworks for sales teams managing $100M+ annual revenue.
  • Automated weekly performance reporting and COVID-19 impact dashboards, reducing manual effort by 20+ hours/week.

Education

2022 → 2023

Technology Leaders Program

Plaksha University

In collaboration with UC Berkeley and Purdue University