Hi, I'm @jinank,
Machine Learning Engineer
who builds self-healing
AI systems that act.
Work Experience
End-to-end ownership through design → architecture → deployment → optimization, collaborating with cross-functional teams to deliver real impact.
SUNY Polytechnic Institute
Graduate Research & Teaching Assistant - ML/NLP
Jan 2025 – May 2026
- •Developed LLM-based NLP pipelines using Phi-3 and instruction-tuned transformer models for document similarity, sentiment analysis, and semantic understanding of Russell 3000 earnings calls (NYBPC 2025 research)
- •Automated Python preprocessing and feature engineering pipelines with pandas, NumPy, and scikit-learn, reducing manual text processing time by 25% across faculty research workflows
- •Built Power BI dashboards and reproducible ML research pipelines supporting academic publications and faculty-led financial analytics projects under Dr. Robert Edgell
- •Mentored 90+ undergraduate students in Python, SQL, machine learning, and applied analytics as Teaching Assistant
- •Defined measurement strategies using precision, recall, F1, ROC-AUC, and statistical significance to guide model decisions
Core Expertise
ML Pipelines & Analytics
Advanced research in financial NLP using transformer-based models and automated data engineering.
- Financial Text Analytics Pipelines
- Reproducible ML Workflows
- Streamlit & Power BI Dashboards
Mentorship & Leadership
Leading technical workshops and mentoring over 90 graduate students in machine learning.
- 90+ Students Mentored
- Technical Workshop Design
- Innovation Advocate
Academic Excellence
Pursuing MS in Data Science with a focus on Deep Learning and Time Series Analysis.
- MS in Data Science (GPA 3.97)
- Outstanding Graduate Student Award
Graduate Research & Teaching Assistant focused on LLM-based financial NLP
Rethinksoft
Senior Software Engineer
Apr 2016 – Jul 2024
Engineered an end-to-end ML prediction system for a Dubai-based client, driving $0 → $1M MRR in year one. Designed conversational workflows with dialog state management, slot-filling, and human-escalation paths; built Node.js/Python bot orchestration services with REST and OAuth2/JWT integrations to internal and third-party systems.
"Exceptional engineer who bridges the gap between ML research and production code."
Openxcell Technolabs
Software Developer
Mar 2013 – Mar 2016
Founding engineer for the Orderhive platform, supporting 1,500+ global merchants and contributing to a $45M acquisition.
Founding Engineering
Built core ecommerce integrations and optimized database performance by 40%.
"Instrumental in scaling our backend and securing our acquisition."
Projects
jinankthakker284 contributions in 2026
GitHub Contribution Grid · github.com/jinank
Leadership & Awards
Recognized for academic excellence and technical contributions in the field of engineering and data science.
- Outstanding Graduate Student Award – Data Science & Analytics (2026)
- Codex Creator Challenge – Handshake AI (2026)
- Program Head @ NYBPC (2025–2026)
- Chairperson @ IEEE LDRP-ITR Student Branch (2010–2012)
for SAs
GNSS Spoofing Detection
Developing transformer-based models for GNSS/GPS spoofing detection using sequential signal patterns and time-series features. Evaluating robustness using recall, F1, ROC-AUC, false-negative rate, and error analysis.
LLM Hyperparameter Optimization
Conducted a 2³ factorial experiment on LLM sampling parameters to quantify impact on readability and output quality. Applied ANOVA to identify top-p as the strongest readability driver (F=4.915, p=0.036). Findings presented in Streamlit.
Patent Phrase Matching
End-to-end DeBERTa semantic matching pipeline for patent phrase similarity across a 10,000-document benchmark. Implemented preprocessing, fine-tuning, validation, dashboards, and reporting - achieving 90%+ accuracy.
Deep Reinforcement Learning for Algorithmic Trading
Multi-agent RL environment for high-frequency trading simulations using FinRL with Q-Learning strategies.
End-to-End MLOps Pipeline
Automated model retraining and deployment with CI/CD on AWS, including validation checks and monitoring.
Knowledge-Graph Enhanced RAG System
Improved LLM context retrieval with structured knowledge extraction using Neo4j and LangChain.
Auto-Scaling Microservices on AWS
Resilient backend architecture with Kubernetes handling peak loads automatically - improving response time by 35%.
RethinkSoft Cloud – AI Agent Platform
Browser-based AI agent SaaS with onboarding, Chrome extension, workflow templates, task history, dashboards, and usage-based pricing. Defined product metrics for activation, workflow execution, task success, retention, and monetization readiness.
RethinkSoft Cloud – AI Agent Platform
Browser-based AI agent SaaS with onboarding, Chrome extension, workflow templates, task history, dashboards, and usage-based pricing. Defined product metrics for activation, workflow execution, task success, retention, and monetization readiness.
Technical Profile
High-Agency EngineerMy approach combines the rigor of traditional software engineering with the flexibility of new AI architectures.
- ›Microservices Architecture: Designing distributed systems with high availability.
- ›Data Engineering: Complex ETL pipelines with 40% performance optimization.
- ›AI Mentorship: Over 90 students mentored in ML and data analytics.
- ›Founding Engineer: Direct contribution to million-dollar acquisitions (Orderhive -> Cin7).
Sharing
Program Head
Startup competition orchestration and technical mentorship.
Coming SoonEducation & Mentorship
Data Science & MLActive mentorship of students and professionals in the AI and data analytics field.
Education
Master of Science in Data Science & Analytics (Machine Learning)
GPA 3.97/4.00 · Utica, NY. Relevant Coursework: Applied Machine Learning, Deep Learning, NLP, Big Data Analytics, Experimental Design, Data Visualization.
Bachelor of Engineering in Computer Engineering
Graduated June 2013 · Gujarat, India. Relevant Coursework: Data Structures & Algorithms, Database Management, Web Development, Java, C, C++.