cv
Résumé PDFGeneral Information
Full Name
Jigar Thummar
Location
Toronto, ON, Canada
Education
2024 – 2025
Master of Professional Studies in Analytics
Northeastern University · Vancouver, Canada
- GPA 4.0 / 4.0
2019 – 2023
B.Tech in Computer Science & Engineering
Parul University · Vadodara, India
- GPA 8.2 / 10
2022
Exchange Semester · Information Security
ITMO University · Saint Petersburg, Russia
- GPA 4.3 / 5
Modern Cryptographic AlgorithmsMachine LearningWeb Software Development
Experience
2026 – Present
Associate Engineer – AI
ZS Associates · Toronto, Canada
- Built LangGraph agents that author CMC regulatory documents for a global pharma client — 50% faster authoring with higher correctness.
- Developed RAG systems and agent tools that ground generation in source material.
- Architected scalable AWS ECS services to isolate compute-heavy authoring and document-mapping workloads.
- Designing an agentic HAQ (Health Authority Questions) feature answering regulator questions post-submission.
2025 – 2026
AI Engineer · Early Employee
VaultAI · Vancouver, Canada
- Engineered an autonomous Due-Diligence Agent synthesizing multi-source research into production-ready PDF reports.
- Built secure RAG pipelines for 100+ confidential documents and a multi-company comparison agent (−60% manual effort).
- Fine-tuned Llama models with QLoRA for a 15% accuracy gain and reduced latency.
2023
Data Science Intern
Gilbert Research Center · Coimbatore, India
- Ran ETL on 10K+ medical X-ray images and trained CNNs in PyTorch to 87% accuracy for early diagnostics.
Skills
AI & Agentic Systems
Multi-Agent SystemsRAGLLM Fine-Tuning (QLoRA)Prompt EngineeringNLPModel Evaluation & Guardrails
Frameworks & Libraries
LangChainLangGraphHugging FacePyTorchTensorFlowscikit-learnXGBoostFastAPI
Languages & Databases
PythonSQLRPostgreSQLMongoDBSupabaseVector DBs
Data Engineering
ETL PipelinesPandasNumPyPySparkHadoopData Normalization
Cloud & MLOps
AWSGCPDocker
Visualization & Analytics
TableauPower BIGrafanaMatplotlib
Projects
2025
EduSense — 🏆 2nd Place, Deloitte AI Hackathon
- AI research assistant with a RAG pipeline structuring unstructured data from 25+ sources, plus an interactive dashboard for proactive, data-driven RFPs.
RAGLLMsDashboard
2025
Real-Time Indoor Air-Quality Analytics
- End-to-end IoT ETL pipeline and forecasting models (ARIMA, XGBoost) predicting CO₂ with 90% accuracy, deployed to a Grafana dashboard with alerting.
IoTARIMAXGBoostGrafana
2024
DDoS Attack Detection in Network Traffic
- Engineered 10+ features from 22M+ traffic records with PySpark and applied unsupervised learning (Isolation Forest, K-Means), flagging 376 high-risk instances across 11 clusters.
PySparkIsolation ForestK-Means