AI / ML Engineer & Graduate Researcher

Harshal
Patel

MS Artificial Intelligence · University of Florida · GPA 3.78

Building production-grade AI systems at the intersection of multimodal learning, explainability, and trustworthy AI. 3.3+ years of industry experience at Accenture and Temenos. CPT/OPT Eligible.

🟢 Open to Summer 2026 Internships PyTorch CUDA / C++ FastAPI Docker / K8s AWS
View Projects GitHub ↗ LinkedIn ↗
3.3+
Years Industry Exp.
3.78
MS GPA @ UF
5+
Production Projects
4
Certifications
01

Education

Master of Science in Artificial Intelligence
University of Florida, Gainesville, FL
📅 Aug 2025 – Jun 2027 📊 GPA: 3.78 / 4.0
Applied Deep Learning · Computer Vision · Machine Learning · Signal Processing · NLP · AI Ethics · Distributed Systems · IoT Systems
B.Tech in Computer Science & Engineering
Visvesvaraya Technological University, Karnataka, India
📅 Aug 2017 – Sep 2021 📊 GPA: 8.8 / 10 (≈3.8/4.0)
Data Structures & Algorithms · Machine Learning · Database Systems · Computer Networks · Operating Systems · Software Engineering
02

Experience

Dec 2024 – Aug 2025
Custom Software Engineering Senior Analyst
Accenture · Client: Dell Technologies · Karnataka, India
  • Authored C++/CUDA kernels for GPU node health validation; exposed to Python via pybind11 bindings — reduced cluster bring-up time by 40% across 50+ node configurations.
  • Implemented OpenCV / scikit-image validation module for HPC rack configuration defect detection, cutting manual inspection effort by 60%.
  • Engineered Python microservices processing 10M+ records/day at 99.9% uptime with retry logic, circuit breakers, and dead-letter queues.
  • Built GitHub Actions CI/CD pipelines with AWS deployments reducing release cycle by 30%.
  • Contributed to Dell Omnia open-source HPC/AI cluster provisioning platform (C++/CUDA modules + pybind11).
C++ / CUDApybind11 OpenCVPython AWSGitHub Actions HPC / Slurm
Sep 2021 – Dec 2024
Software Engineer
Temenos · Karnataka, India
  • Designed ETL pipelines ingesting 100K+ banking transactions/day into PostgreSQL and MongoDB with schema validation and idempotency.
  • Reduced REST API p95 latency by 60% (to sub-200ms) via composite index redesign and Redis caching.
  • Re-architected batch jobs into RabbitMQ event-driven microservices achieving 3× throughput.
  • Implemented ML-based fraud detection module (sub-50ms SLA) reducing false-positive rate by 22%.
PythonPostgreSQL MongoDBRedis RabbitMQML / Fraud Detection REST APIs
03

Projects

🔍
FraudLens — MultiModal Fraud Detection
Production-grade fraud detection pipeline fusing SigLIP 2 (vision), DistilBERT (NLP), and structured transaction analysis via learned cross-modal attention. Includes Captum XAI explainability with per-modality integrated gradients.
SigLIP 2 + DistilBERT Cross-Modal Attention Captum XAI FastAPI Docker
PyTorchHugging Face FastAPIDocker
🛡️
TrustCheckAI — ML Bias & Compliance Monitor
Real-time ML bias and compliance monitoring platform with Prometheus/Grafana dashboards tracking fairness and SLA compliance for 1,000+ active users. Reduces MTTR by 45% via automated alerting.
1,000+ users 10K+ events/day 45% ↓ MTTR
PrometheusGrafana StreamlitLIME Slack
👗
Fashion Aura — AI Virtual Try-On
Computer vision pipeline for garment segmentation across a 78,000-item catalog with 95% accuracy. Powered by Gemini 2.0 Flash multimodal model with async handling of 1,000+ concurrent requests sub-3s.
78K+ catalog items 95% accuracy 1K+ concurrent req.
OpenCVGemini 2.0 Flash FastAPIAWS S3 Nginx
🎫
Support Ticket System — NLP Auto-Triage
Fine-tuned DistilBERT NLP classifier achieving 91% accuracy across 8 ticket categories. Eliminated 70% manual triage via Celery async queues, sustaining 10,000+ API requests/day.
91% accuracy 10K+ req/day 70% triage reduced
DjangoDistilBERT PostgreSQLRedis Celery
04

Technical Skills

Languages
Python C++ Java JavaScript SQL
AI / ML
PyTorch TensorFlow scikit-learn Hugging Face OpenCV Grad-CAM Captum XAI
HPC & GPU
CUDA pybind11 MPI OpenHPC Slurm
Backend & Cloud
FastAPI Flask Django AWS Docker Kubernetes CI/CD
Databases & Systems
PostgreSQL MongoDB Redis Kafka MySQL Linux
Monitoring & Ops
Prometheus Grafana GitHub Actions Postman MATLAB
05

Certifications

☁️
AWS Cloud Technical Essentials
Amazon Web Services
🔷
Azure Fundamentals (AZ-900)
Microsoft
⚙️
Intro to Kubernetes (KCNA)
Cloud Native Foundation
🧠
Deep Learning Specialization
NVIDIA


🌐
Open Source · Dell Omnia
Contributor to Omnia by Dell Technologies — open-source HPC/AI cluster provisioning tool. Contributed C++/CUDA algorithm modules with pybind11 Python bindings, GPU node automation scripts, and an image-based configuration validation module.
Get In Touch

Let's Build
Something Together

Open to Summer 2026 internship opportunities in AI/ML Engineering & Software Engineering. Let's connect.

✉️ harshal27patel@gmail.com 📞 +1-352-328-6754 ⚡ GitHub 💼 LinkedIn
Available for Summer 2026 Internships