Shubham Deshmukh

Software Engineer Intern @ J. Craig Venter Institute

About

Hey! My name is Shubham Deshmukh. I have 2+ years of experience in backend systems and AI pipelines. I am currently looking for Software Engineer/ AI Engineering roles. I build production-ready AI systems that turn unstructured data into reliable, decision-ready intelligence. I’m a Software Engineer (AI/ML) with an MS in Computer Science from Virginia Tech (May 2025), specializing in agentic AI workflows, LLM systems, and scalable ML deployment. At J. Craig Venter Institute, I architected and deployed agentic AI pipelines using LangChain to automate large-scale scientific literature mining. This work supports the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) — a NIH/NIAID-funded national bioinformatics resource used by the global infectious disease research community. I built multi-step LLM workflows with tool orchestration, structured output validation, retry/backoff logic, and human-in-the-loop review systems. These pipelines integrated multiple LLM providers, improved ingestion throughput by 3×, increased extraction F1 by 35%, and reduced expert review time by 40%. In parallel, as a Graduate Research Assistant at Commonwealth Cyber Initiative, I developed end-to-end computer vision systems across 500k+ environmental images. I built OpenCV + U-Net segmentation pipelines, trained CNN and Vision Transformer models in PyTorch/CUDA (R² = 0.898), and deployed inference services using Streamlit and Docker to reduce latency and improve usability. Previously at Wipro, I built and maintained backend and cloud-based systems using ASP.NET Core and Azure, strengthening my foundation in scalable architecture, reliability, and production-grade engineering. What sets me apart: 1. Production agentic AI systems using LangChain 2. LLM pipelines with validation, tool orchestration & human-in-the-loop design 3. Strong backend + cloud engineering fundamentals 4. Proven ability to take systems from data → model → deployment → measurable impact I’m seeking Software Engineer (AI/ML) or AI Engineer roles where I can build intelligent systems that operate reliably at scale.

Country

United States

City

San Francisco Bay Area

Industry

Computer Software

Skill

Agentic Workflows, Amazon Web Services (AWS), AWS Lambda, API Development, Large Language Models (LLM), HTML, Cascading Style Sheets (CSS), JavaScript, React.js, D3.js, GitHub, Java, UI, SQL, Artificial Intelligence (AI), Electronics, German, Geography, Software Project Management, Web Development

Experience

Software Engineer Intern

J. Craig Venter Institute

2025-7 - Present · 1 yr 3 mos

Rockville, Maryland, United States

Designed Agentic AI based literature-mining pipeline using Gemini cutting bioinformatician review time by 40%. • Engineered NCBI E-utilities ingestion with rate-limits, and retries, increasing PubMed full-text ingestion throughput 3×. • Implemented Streamlit human-in-the-loop UI with prompt versioning and edits, improving extraction F1 by 35% overall.

Software Engineer Intern

Commonwealth Cyber Initiative (CCI)

2024-8 - 2025-9 · 1 yr 2 mos

Arlington, Virginia, United States

Trained and optimized CNN and Vision Transformer models in PyTorch/TensorFlow with CUDA, achieving 89% test accuracy and improving CDOM prediction performance by 30%. • Curated USGS HIVIS/NWIS dataset: 500k+ RGB images, 111 sites, and six optically active parameters labeled. • Developed a novel OpenCV pipeline with U-Net water segmentation model, astral day/night filtering, and 20% water-pixel threshold to enforce consistent water-pixel quality. • Deployed inference service using Streamlit/Docker, reducing end-to-end prediction latency from 2.3s to 1.6s per image.

Software Engineer Intern

J. Craig Venter Institute

2024-6 - 2024-8 · 3 mos

Rockville, Maryland, United States

Delivered interactive React + D3.js dashboards for HSP 3.0, enabling drilldowns for 20+ researchers weekly. • Containerized Python/R ETL pipeline with Docker and AWS Lambda/S3, reducing manual data prep steps by 70%. • Optimized backend API queries and indexing, cutting end-to-end batch processing latency by 20% per run.

Software Engineer

Wipro

2022-5 - 2023-6 · 1 yr 2 mos

Pune, Maharashtra, India

Built and maintained data-driven backend services across 10+ ASP.NET Core applications, improving service efficiency by 25% and enabling scalable analytics workflows. • Automated Tableau Server permissions, refresh schedules, reducing report turnaround time by 15% monthly. • Resolved 40+ production POS incidents, improving system uptime by 20% and accelerating transaction throughput.

Software Engineer Intern

ifm

2021-8 - 2022-1 · 6 mos

Pune, Maharashtra, India

Developed real-time patient monitoring system using YOLOv4-tiny and Mediapipe, achieving 98.82% pose classification accuracy. • Integrated CV/Deep Learning models into Flask web app with alerts, reducing manual observation workload by 30%. • Implemented Tkinter GUI with pose logs and alerts, exporting CSV reports for nurses and doctors daily.

Education

Virginia Tech

Computer Science

2023-8 - 2025-7 · 2 yrs

Artificial Intelligence and Data Analytics Specialization Relevant Coursework: Intermediate Data Structures and Algorithms Analysis, Web Application Development, Software Engineering, Introduction to Deep Learning, Introduction to Urban Computing, Machine Learning with Big Data, Information Visualization

Vishwakarma Institute Of Technology

Electronics and Telecommunication Engineering

2018-8 - 2022-7 · 4 yrs

Artificial Intelligence and Data Analytics Specialization Relevant coursework: Data Structure and Algorithms, C Programming, Python Programming, Computer Vision, Signals & Systems, Digital Signal Processing, Natural Language Processing, Cloud Computing Activites and societies: E - Waste Collection Drive, Technical Volunteer for "Wings of Fire" in Vishwakarandak'18 and Melange'19.