website logo

Harsh PV

Full-stack Software Engineer | AI & Backend | Seeking Full-time Roles May 2025

Masters in Computer Science @ UIUC. Experienced in AI, backend, distributed systems, and cloud. Open to full-time SWE roles starting May 2025.

Get in Touch

Work Experience

University of Illinois - Urbana Champaign

Software Engineer

Jan 2024 — Current
  • Collaborated with Prof. Rohit Bhargava and Mayo Clinic to optimize model inference times using quantization and pruning, integrating NVIDIA TensorRT, Triton, RAPIDS, and DALI, achieving a 96% reduction in inference times (6 hours to 2 minutes).
  • Worked with Prof. Ahmed Elbanna and USC on a $5M NSF-funded project, fine-tuning Vision Language Models (VLMs) for structural engineering defect identification.
  • Collaborated with Dr. Huang at SALT Lab to address hallucination and factuality loss in LLMs using knowledge graph-based RAG systems.

Cisco Systems

Software Engineer - Backend

Aug 2021 — Jan 2024
  • Developed features for Cisco’s configuration translation and migration engine, resulting in a 30% increase in product adoption.
  • Spearheaded core DB scalability via partitioning and reduced CPU utilization from 60% to 25% through query optimizations.
  • Automated network deployments using Terraform and Ansible, reducing deployment times by 60%.

Novartis AG

Software Engineering Intern

Jan 2021 — July 2021
  • Designed and implemented an ETL-driven KPI dashboard for ITSM data, improving team efficiency by 30%.
  • Integrated a machine learning-powered capacity planning service using XGBoost to benchmark 25+ project and team KPIs.

Siam Computing

NLP Engineering Intern

May 2019 — Aug 2019
  • Developed RASA-based chatbot with complex dialog routes, reducing contact center interactions by 40%.
  • Built a facial recognition-based attendance system using facial encodings, deployed on Raspberry Pi.

Vellore Institute of Technology

Research Assistant

Mar 2019 — Jan 2021
  • Developed an image captioning system for surveillance using Xception CNN and LSTM, trained on Flickr8K.
  • Implemented rotation-invariant CNN for image augmentation, enhancing CCTV accuracy.

Featured Projects

Clarity

Q&A knowledge bot using Pytorch, Streamlit, Langchain, OpenAI, HuggingFace, BS4, Bootstrap. Extracts and indexes information for rapid query resolution, reducing turnaround time for customer-facing professionals.

Source Preview
Clarity

NetBot

Network automation tool using Pytorch, Django, Langchain, HuggingFace, Bootstrap. Fine-tuned LLaMA 3.1 7B model on Cisco device commands, streamlining network management and reducing manual configuration.

Source Preview
NetBot

Police Surveillance Assistant

AI-powered surveillance system (Python, Flask, React, GCP, CNN, LSTM, CoCo, MongoDB, OpenCV, Dialogflow). Real-time video analysis, facial/object detection, automated police alerts, and chatbot interface.

Source Preview
Police Surveillance Assistant

Personal Assistant for Blind

AI-powered device for visually impaired (Python, Flask, React, GCP, Docker, OpenCV, Dialogflow). Real-time friend detection, currency recognition, OCR-based image-to-text, distributed camera feeds, and voice commands.

Source Preview
Personal Assistant for Blind

DA - Developer Assistant

Quick and hassle-free development tool (Python, Flask, Dialogflow, Git) for project setup, source control, and auto debugging.

Source Preview
DA - Developer Assistant

Achievements & Publications

Achievements

  • Winner, UIUC NCSA Ashby Hackathon (multi-agent code LLM using GPT-4 & Claude)
  • Cisco FY23 Achievers Award
  • Cisco Excellence Awards (2021, 2022, 2023)
  • Winner, SIH 2020 (Govt. of India)
  • Runners up IBM Hack the COVID
  • Winners Parallax Hackathon
  • Winner JU McAfee Hackathon

Publications & Patents

  • "Captioning for Motion Detection in Video Surveillance Applications Using Deep Learning" (IJITEE)
  • DeveAid Patent (201941053665): Developer assistant tool for automating project environment setup.

About Me

Harsh PV is a full-stack Software Engineer (UIUC MSCS '25) with experience in AI, backend, distributed systems, and cloud. Technical Skills: Languages: C, C++, Python, C#, Java, MySQL, Neo4J, JavaScript, Typescript, HTML, CSS, Git, Shell, Golang Frameworks/Tools: Django, HBase, Flask, Langgraph, Docker, Kubernetes, MongoDB, GraphQL, AWS, Kafka, Terraform, Linux, Jenkins, PostgreSQL, Redis, Pinecone, FlowAI Concepts: Data structures and Algorithms, RAG, Operating Systems, Network Protocols, Distributed Systems, TCP/IP stack, MVC, Agile, Scrum, DevOps Education: - University of Illinois - Urbana Champaign, MS in Computer Science, Jan 2024 – May 2025 - Vellore Institute of Technology, BTech in Computer Science, Aug 2017 – May 2021

Harsh PV