Skip to content
Sahil HamalResume

Reston, VA · DC Metro

Sahil Hamal

AI Engineer · LLM & Agentic Systems · Full Stack

Software engineer specializing in LLM integration, RAG systems, and full-stack AI features. Currently at American Express. Previously at Johns Hopkins (AstroPath, published in Science), Deutsche Bank, and Zoom.

About

Engineer focused on getting AI features into production safely.

I'm a software engineer with 4+ years of production experience, specializing in LLM integration, RAG systems, and full-stack AI features. I currently work at American Express, where I've shipped retrieval-augmented pipelines and LLM-powered classification systems that moved measurable accuracy and latency metrics in production.

My day-to-day spans LLM integration (OpenAI API, LLaMA, Claude), RAG pipelines, semantic search, vector embeddings, prompt engineering, LLM evaluation, and agentic workflows (ReAct, LangChain). On the full-stack side I work in React, Vue.js, Spring Boot, Node.js, Python, Java, and TypeScript, deployed on AWS (Lambda, S3, API Gateway), Docker, and Kubernetes.

Before American Express, I was a published contributor to AstroPath — an NIH-funded cancer imaging platform published in Science (2021), covered by The Economist, and recipient of the 2021 Falling Walls Life Sciences Award. My graduate research at Virginia Tech under Dr. Chris North focused on explainable AI, and earned the Torgersen Outstanding Graduate Research Award. That work still shapes how I approach LLM evaluation and failure-mode analysis today.

Based in Reston, VA. Open to AI Engineer, Senior SWE, and Full-Stack Engineer roles — hybrid or remote.

Experience

Selected roles.

  1. June 2023 – Present

    Washington, DC

    Software Engineer II, AI & Full Stack · American Express

    • Designed and shipped an LLM-powered transaction categorization system (OpenAI API + embedding-based retrieval for few-shot context) that replaced a legacy rules-based pipeline, improving classification accuracy from 78% → 94% on a held-out production dataset of ~50K labeled transactions.
    • Built a RAG pipeline over internal transaction metadata using LLaMA embeddings + ElasticSearch as the vector/keyword hybrid store, served via Spring Boot; reduced p95 query latency ~40% through Redis caching and re-ranking optimization.
    • Established LLM evaluation framework with precision/recall tracking, prompt regression tests, and A/B comparisons across model versions; used telemetry to iteratively tune prompts and retrieval parameters based on measurable impact.
    • Optimized LLM inference tradeoffs (model choice, token budget, batching, caching) to balance latency vs. accuracy, cutting per-request cost while maintaining classification precision above SLA.
    • Led design reviews and code reviews across a 3-engineer team, driving production rollout decisions, architecture proposals, and safe deployment patterns for AI features in a regulated financial environment.
    • Partnered cross-functionally with Product and Analytics to translate business requirements into AI features; resolved ingestion/search inconsistencies with idempotent pipelines and validation checks.
  2. May – Aug 2022

    San Jose, CA

    Full Stack Developer Intern · Zoom Video Communications

    • Built a full-stack video archive for Zoom Events with tag-based ElasticSearch retrieval and real-time indexing across thousands of event recordings.
    • Architected hybrid AWS S3 + MongoDB storage with CDN delivery and an ElasticSearch metadata pipeline enabling sub-second full-text search, tag filtering, and playback.
  3. Oct 2019 – June 2021

    Baltimore, MD

    Software Developer, AstroPath (NIH) · Johns Hopkins University

    • Application developer for AstroPath — an NIH-funded cancer imaging platform published in Science (June 2021), covered by The Economist, and recipient of the 2021 Falling Walls Life Sciences Award.
    • Built the full-stack cell-view annotation and cancer detection interface used by JHU immunologists under NIH data governance and IRB compliance.
    • Contributions supported securing 5+ years of continued NIH research funding.
  4. June 2018 – Oct 2019

    Cary, NC

    Software Developer · Deutsche Bank

    • Built a real-time dependency tracking system for financial communication platforms with automated notifications, preventing compatibility failures across enterprise systems.
    • Designed and implemented an automated Ansible-based deployment pipeline for financial applications into secured, compliance-governed banking environments with full audit trail and rollback.
  5. June – July 2017

    Dallas, TX

    Software Developer Intern · PricewaterhouseCoopers (PwC)

    • Built a financial account management application for PwC Tax & Technology with report generation, delivered within a Big Four professional services environment.

Projects

Selected work.

CVE Triage Agent

In Progress

LangChain-based ReAct agent that ingests CVE descriptions and vendor advisories, classifies severity and affected components, and drafts structured remediation notes. Built to bridge AI engineering with offensive-security domain knowledge.

PythonLangChainReActLLM APIs

AstroPath — NIH Cancer Imaging Platform

Application developer for an NIH-funded cancer imaging platform that maps immune cell populations in tumor microenvironments from multiplex immunofluorescence tissue images.

  • Published in Science (June 2021)
  • Covered by The Economist
  • 2021 Falling Walls Life Sciences Award
Full-stack web appImage processing pipelinesNIH data governance

RAG-Powered Document Intelligence

Retrieval-augmented generation systems enabling natural language querying over large document corpora with hybrid vector/keyword search. Implemented semantic search with vector embeddings, LLM integration, and context-aware response generation.

Next.jsTypeScriptPythonPineconeElasticSearchGoogle Gemini API

Explainable AI Research (MS Thesis)

Gradient-based visualization techniques for interpreting dimensionality reductions (UMAP, t-SNE). Informs my current approach to LLM evaluation and failure-mode analysis.

  • Torgersen Outstanding Graduate Research Award
PythonPyTorchUMAPt-SNESHAP/LIME

Publications

Peer-reviewed work.

“Multiplexed imaging of immune markers and tumor tissue architecture in human cancer”

Science · June 2021

Awards

Recognition.

  • Falling Walls Award — Life Sciences

    2021 · for AstroPath

  • Torgersen Outstanding Graduate Research Award

    Virginia Tech

Technical Skills

Tools I reach for.

AI / GenAI

LLM integration (OpenAI API, LLaMA, Claude)RAG pipelinesSemantic searchVector embeddingsPrompt engineeringLLM evaluationHugging Face TransformersPyTorchAgentic workflows (ReAct, LangChain)

Languages

PythonJavaJavaScriptTypeScriptSQLC#

Backend / APIs

Spring BootNode.jsRESTGraphQLMicroservices

Frontend

ReactVue.jsReact NativeAngularHTML5CSS

Data / Search

ElasticSearchRedisVector storesMongoDBPostgreSQLMySQL

Cloud / DevOps

AWS (Lambda, S3, API Gateway, Bedrock-adjacent workflows)DockerKubernetesAnsibleCI/CDGitHub Actions

Education

Academic background.

Virginia Tech

M.S., Computer Science

Blacksburg, VA · 2021 – 2023

  • Concentration: Visual Analytics & Explainable AI
  • Advisor: Dr. Chris North
  • Thesis: Interpreting Dimension Reductions Through Gradient Visualization
  • Award: Torgersen Outstanding Graduate Research Award

Troy University

B.S., Computer Science

Troy, AL · 2014 – 2018

  • President, Computer Science Club
  • University Ambassador