Gen AI Developer & Freelance Consultant with a love for LLMs, agentic AI & scalable systems
// 04. about
Hi, I'm Muskan - a Gen AI Developer & Freelance Consultant who builds intelligent systems that actually work. Passionate about LLMs, agentic AI and turning complex data into impactful solutions.
Focused on Python, LangChain, FastAPI and RAG pipelines - with a mathematics foundation from Simon Fraser University and real-world experience across Canada and India.
// 01. education
// 05. skills
// 03. projects
A selection of AI, machine learning, software engineering, data, research and robotics projects.
Agentic resume-to-job matching system that parses resumes and job descriptions, evaluates skill alignment, generates hiring recommendations, and applies deterministic human-review guardrails for safer decisions.
Transformer-based fake news classification system comparing BERT, RoBERTa, LSTM and SVM models. Achieved 99.73% accuracy with RoBERTa and incorporated explainability using Integrated Gradients and Attention Rollout.
Production-style FastAPI telemetry gateway for local environmental monitoring, handling duplicate events, device restarts, clock skew, transactional state updates, WebSocket reliability and focused failure testing.
Linear programming model optimizing emergency blood distribution across hospitals in Northern BC while supporting allocation decisions under demand uncertainty.
Mathematical and stochastic modeling project exploring species diversification under the Red Queen Hypothesis using deterministic ODEs and Gillespie simulations.
Machine learning model predicting walking, cycling and bus transportation modes from time-series accelerometer data using feature engineering and classification.
Formal CNN robustness verification system using interval arithmetic and Lipschitz bounds to analyze neural network behavior under perturbations.
Stock trading simulator using real market API data, order matching, portfolio tracking and object-oriented software architecture.
Autonomous Raspberry Pi robot using sensor fusion and PID control logic to detect and follow a predefined path.
Desktop point-of-sale application supporting product categories, tax calculations, receipt generation and customer feedback workflows.
// 03. experience
Researching and developing AI/ML solutions for flight-object classification, trajectory prediction, sensor fusion and state estimation across 5+ aerospace research directions, experimenting with PINNs, CNN, LSTM, GRU, Random Forest & XGBoost.
Processing and analyzing 50K+ data points using Python, feature engineering, statistical/time-series analysis, model experimentation, LoRA/QLoRA fine-tuning and validation to optimize ML architectures for flight-data applications.
Translating AI/ML research into practical AI solutions and automated workflows, supporting model evaluation, fine-tuning, trajectory analysis and data processing while collaborating across technical and operational teams.
Developed LLM-powered medical using Indian medical databases and insurance AI solutions using Python/C#, processing policy and claims data with OCR, Mistral fallback, RAG, SQL, vector databases, prompt engineering, AI evaluation and Responsible AI guardrails to detect the frauds in the documents.
Built AI agents and automated workflows across 8+ healthcare, insurance, CRM, ERP, inventory and fraud solutions using Python, C#, R, REST APIs, Azure, AWS, PySpark and Databricks.
Designed RAG/NLP pipelines for structured and unstructured data using embeddings, semantic search and vector databases; applied MLOps practices including Docker, Kubernetes, model validation, monitoring and testing, reducing manual effort by 40%.
Applied linear programming and mathematical optimization to improve emergency blood distribution across BC hospitals. Developed quantitative models to support allocation decisions under demand uncertainty and evaluated stochastic scenarios to reduce allocation risk and improve resource utilization.
Supported 60+ students in Python programming and data structures courses through laboratory sessions, office hours, technical guidance and grading. Helped students understand programming fundamentals, algorithms, debugging and problem-solving while receiving consistently strong feedback.
Managed daily, shift-end and month-end reporting across two locations,
analyzing sales, inventory, invoices, cash deposits, purchasing, customer
satisfaction, compliance and operational data using Excel and QuickBooks.
Maintained accurate financial and operational records while monitoring
performance against business targets and identifying discrepancies.
Analyzed promotions, loyalty and reward programs, customer surveys and
historical sales trends to identify purchasing patterns and improve
inventory and ordering decisions. Used operational insights to support
store performance, customer satisfaction and cost efficiency, contributing
to a top-ranked store position and earning a promotion.
Trained and supported team members across daily operations, reporting,
customer service and compliance while helping maintain consistent processes
across multiple locations.
// 04. certifications
Duke University
Coursework covering Java programming, software engineering fundamentals, object-oriented programming, data structures and application development.
Show Credential ↗University of Colorado Boulder
Coursework focused on database concepts, SQL, relational databases, data management and database techniques for data science applications.
Professional Training
Comprehensive data analytics training covering fundamentals through advanced analysis, data cleaning, visualization, SQL, statistics and practical analytical workflows.
Professional Training
Hands-on training in big data engineering, data pipelines, distributed processing and cloud technologies with practical exposure to Google Cloud Platform and Microsoft Azure.
Udemy
Training in algorithmic trading concepts, Python-based quantitative analysis, strategy development, backtesting and financial data analysis.
Udemy
Practical machine learning training covering supervised and unsupervised learning, model development, feature engineering, evaluation and applied ML workflows.
// let's discuss!
Looking for Gen AI & software engineering roles. Project ideas, consulting or just want to talk AI - reach out.
☕ "Let's grab a virtual coffee and build something that matters."
// that's me!
My gaze follows your every move — grab a coffee & watch me smile! ☕