A
Adam Brown
About
Detail
Senior AI Engineer
Madagascar
Résumé
Jobs
verified_user
0% verified
- DSenior AI EngineerDraftKingsNov 2022 - Current (3 years 9 months)Led LLM experimentation and fine-tuning for natural-language match commentary and analytics using GPT-style models (instruction tuning, RLHF workflows). Designed and deployed RAG-based retrieval + generation pipelines for contextualized sports insights, integrating FAISS vector store and semantic embeddings. Built agentic orchestration for multi-step inference tasks using LangChain and custom tool-plugins to combine retrieval, calculators, and rule engines. Implemented prompt engineering, prompt templates, and dynamic context windows to reduce hallucinations and improve factuality in generated outputs. Migrated model serving to GPU-accelerated inference endpoints, reducing latency 35% and enabling sub-second responses for high-throughput AP
- KAI EngineerKari Technologies International (KTI)May 2018 - Oct 2022 (4 years 6 months)Developed transformer-based NLP models (BERT, GPT-2/3) for intent classification, summarization, and semantic search across customer interactions. Engineered domain-specific embedding pipelines and vector search layers using FAISS and custom ANN for low-latency retrieval. Built scalable feature stores and data ingestion pipelines to support supervised learning and embedding generation at scale. Operationalized model deployments with Docker and Kubernetes, integrated CI/CD and automated model validation checks. Implemented explainability dashboards and SHAP-based feature attribution for stakeholder-facing reports. Designed reinforcement-learning prototypes for recommendation and personalization use cases tested in production-like environment
Machine Learning EngineerInnodata Inc.Jul 2015 - Apr 2018 (2 years 10 months)Built high-precision supervised models for document classification and information extraction, achieving 95%+ precision on large corpora. Implemented semantic search and retrieval systems using sentence embeddings and FAISS for fast lookup across enterprise documents. Designed data augmentation pipelines using VAEs and GANs to expand training datasets for low-resource classes. Engineered ETL and preprocessing for noisy OCR and PDF inputs, improving downstream extraction accuracy. Deployed serverless inference endpoints and batch scoring with AWS Lambda and managed scheduling to meet SLAs. Collaborated with clients to integrate models into editorial and legal workflows, delivering automation and measurable efficiency gains.
Education
verified_user
0% verified
- UBachelor's degree, Computer ScienceUniversity of Nevada, Las Vegas (UNLV)Aug 2011 - May 2015 (3 years 10 months)
Projects (professional or personal)
verified_user
0% verified
- VVariance Analysis AgentJan 2024 - Current (2 years 7 months)Developed agentic system that detects and explains financial variances by combining structured ledger data ingestion, anomaly detection models, and LLM-generated human-readable explanations.
- FFinancial Document UnderstandingJan 2023 - Jan 2025 (2 years 1 month)Built end-to-end RAG pipeline for P&L and GL document analysis using Azure Document Intelligence for layout/OCR, embeddings stored in Chroma, LangChain agents for multi-step question-answering, and Azure ML for model lifecycle.
- EEmbedding & Vector Store MigrationJan 2022 - Current (4 years 7 months)Migrated semantic search from on-disk FAISS to managed Pinecone, improving cluster resilience and reducing query latency under load.
- EExplainability FrameworkJan 2021 - Jan 2023 (2 years 1 month)Deployed SHAP-based dashboards and integrated with monitoring to present model explanations to non-technical stakeholders.