Resume

Vedaang Chopra

Software Development Engineer (ML/AI)
Atlanta, GA | +1 404-740-9905 | vedaangchopra@gatech.edu | LinkedIn | GitHub

Data & Pre-Modelling: Python (NumPy, Pandas), OpenCV, Feature Engineering, Data Augmentation, Plotly/Matplotlib, Vector Databases (FAISS, Redis), Elasticsearch
Representation & Modelling: PyTorch, TensorFlow, Hugging Face Transformers/Diffusers, LangChain, LangGraph, PyTorch Geometric, scikit-learn, CNNs, RNNs/LSTMs, Transformers, GNNs, RL, Generative & Agentic AI
Optimization & Efficiency: Quantization, Pruning, Knowledge Distillation, LoRA/QLoRA, ONNX Runtime, OpenVINO, Mixed Precision, Distributed Training (PyTorch Distributed)
Inference & Serving: vLLM, MLflow, FastAPI, Docker, Kubernetes, REST APIs, Latency Profiling, Model Monitoring
Systems & Infrastructure: Go, C/C++, PostgreSQL, Redis, RabbitMQ, Airflow, Bash, Git, Linux, High-Performance Computing (Slurm, tmux), Edge/Cloud Deployment

Fortinet Technologies Inc. (AIOps, R&D) | Software Development Engineer I & II (ML/AI) | Bengaluru, India | Feb 2021 – Jul 2025

Led development of an Agentic RAG chatbot (LLMs + tool calling) — a Fortinet Hackathon 2023 (5th place) prototype that evolved into a production feature, reducing issue-resolution time by over 70%.
Built an unsupervised connectivity-threshold model (PCT/172022/958026) to detect wireless anomalies, cutting manual troubleshooting efforts by more than 75%.
Implemented DBSCAN-based anomaly detection for SD-WAN telemetry, proactively preventing over 50% of potential network outages.
Re-architected OpenSearch ingestion pipelines using async I/O and Golang, scaling event throughput from under 50 to over 2000 events per second.
Deployed edge-optimized ML models via ONNX Runtime, lowering inference latency by 40% across distributed FortiGate devices.
Automated multi-domain data collection and SLA forecasting for 60+ ML classifiers and 4 SLA categories, enabling 7-day AI-based performance prediction.

Georgia Institute of Technology | M.S. in Computer Science (Machine Learning Specialization) | Atlanta, GA | Aug 2024 – Dec 2026

Maharaja Surajmal Institute of Technology (GGSIPU) | B.Tech. in Information Technology | New Delhi, India | Aug 2016 – Aug 2020

Which–VLM Router | CS 8803 Systems for AI (Advisor: Dr. Anand Iyer) | Aug 2025 – Present

Designing a semantic router that dynamically directs multimodal (text, vision, audio) queries across multiple VLM/LLM endpoints using budget–aware and retrieval–augmented policies.
Expected to achieve 30% lower inference cost with comparable accuracy to Mixture-of-Experts baselines.

Edge Glass Assistant | CS 8803 VLM & LLM (Advisors: Dr. Zsolt Kira, Dr. Alan Ritter) | Aug 2025 – Present

Building a lightweight multimodal assistant aligning vision, audio, and text embeddings through frozen encoders and quantized projectors for on–device reasoning.
Aims to deliver 2× faster inference with minimal accuracy degradation compared to existing edge–ready frameworks.

ATHENA | CS 8903 Agentic AI (Advisor: Dr. Vijay Madisetti) | Jan 2025 – May 2025

Developed ATHENA, a multi-agent generative AI framework unifying LLMs, diffusion models, and memory orchestration for automated screenplay-to-video generation.
Outperformed prior single-agent baselines (BLEU-4 +18%, CLIPScore +22%).

Ontology-based Text Classification | Undergraduate Research (Advisor: Dr. Sonika Malik) | 2019 – 2020

Proposed a semantic ontology framework using the Human Disease Ontology (DOID) for biomedical text classification.
Improved accuracy by 10% over classical ML baselines (Published in CEUR Workshop Proceedings Vol. 2786).