About
Me.
Hi, I'm Janmejay but everyone calls me Jam. I'm a builder passionate about creating products people love.
I'm passionate about architecting systems, innovating features, and providing real value to people. Every day I think about problems I can solve in the world, and tools like Cursor and Claude Code help me build solutions faster than ever before. I'm drawn to complex technical challenges that lead to measurable product outcomes.
I founded CosmiQ, a spirituality platform serving over 1,000 users with 10k readings per month. We are on track to complete over 1 million monthly readings by the end of 2026. I designed everything from the backend architecture to the intuitive interface that brings it to life.
When I'm not building, I stay active through men's league soccer and gym sessions, and I find my balance through a daily meditation practice.
Recent Accomplishments.
Research & Publication
2nd Author
"What patients ask: Evaluating google search queries and webpage transparency in common ophthalmologic malignancies"
The Engineering Behind CosmiQ
CosmiQ is the best showcase of my software engineering skills. I talk about my accomplishments, problems solved, and more.
Read Technical Deep DiveEducation
Formal training in computer science and software engineering.
University of Colorado Boulder
Master of Science in Computer Science
AI Graduate Certificate
March 2024 – June 2025
3.98/4.0 GPA
Western University
Bachelor of Engineering Science in Software Engineering
September 2018 – December 2022
Work Experience
Building products, scaling platforms, and delivering measurable impact.
Jul 2025 – Present
Architected and developed a spirituality platform serving 1000+ users with a 6.5% conversion rate to premium tier, generating 10,000+ monthly readings, built with TypeScript, Next.js, Supabase, Inngest and hosted on Vercel.
Reduced LLM token consumption and operational costs by 68% by developing a sawtooth context compression algorithm that generates conversation summaries every 10 messages.
Developed Conversation Memories, a RAG-based feature with 89% user adoption, using pgvector embeddings and cosine similarity to fetch similar past readings and filter them for relevance in the prompt.
Improved UI responsiveness by 40% through techniques like lazy loading chat history and Zustand-powered optimistic state updates that render changes before server confirmation.
Built analytics and feedback endpoints to quantify total users, MAU, and WAU, alongside retention cohorts and upgrade likelihood scoring, driving product decisions that increased premium tier conversion by 20%.
Ry Consulting
Toronto, ONSoftware Engineering Consultant, Contract Full-time
January 2025 – July 2025
Collaborated with 5 stakeholders to deliver an Azure PoC, modernizing legacy data processing workflows by designing a cloud-native architecture for OneVest API ingestion and Salesforce lead automation.
Engineered ETL pipelines using ADF Dataflows to ingest over 1 million records for over 900K users, handling GraphQL pagination and flattening nested JSON into partitioned CSVs stored in Azure Blob Storage.
Coded Azure Python functions to automate Lead-to-Account conversion by processing leads from Azure SQL DB, reducing manual Sales Operations workload by 40%.
Led formal knowledge transfer sprint, producing technical architecture diagrams and data flow documentation detailing OneVest-Salesforce integration logic to ensure long-term maintainability by internal teams.
January 2024 – December 2024
Owned software roadmap and shipped revenue-critical features like ML models, chrome extension, website and internal tools, contributing to ARR growth from $100K to $600K over 12 months.
Engineered ML models in Python to forecast real estate market trends, generating insights used in 50+ client presentations and driving $100K+ in sales.
Designed and implemented a Chrome extension for realestate.com.au in JavaScript and CSS with more than 200 active users generating $50K+ in revenue.
Translated Figma designs into a conversion-optimized Webflow site, running A/B tests that increased user engagement by 75%.
Projects
Personal projects demonstrating full-stack development, AI/ML engineering, and product thinking.
Intelligent Lifting
Fitness & Wellness Platform
Growth & Impact
- • Scaled to $5,000+ revenue via digital product sales.
- • Combatted "cookie-cutter" routines with data-driven, science-backed plans.
- • Solved accessibility issues by offering excel files with formulas which help users adapt to their MAV.
Technical Implementation
- Built a deployment pipeline with AWS Amplify and CodePipeline to test development branches before pushing to production.
- Developed adaptive training sheets using MEV/MRV principles to dynamically optimize volume toward MAV based on individual recovery.
Agent Monitoring
Microservice
Built a dual-service microservice to ingest and monitor agent events, using Groq (Llama) for realistic simulation and AWS EC2 for containerized hosting.
- Dual-Service StackService A (Ingest API) & Service B (LLM Simulator) linked via Docker Compose.
- Rolling WindowStrict 5-minute/30-event memory buffer with Zod validation for reliability.
Phishing Detection
ML Research Framework
- Comparative study of Logistic Regression vs. Multinomial Naive Bayes
- URL pattern recognition using NLTK (RegexpTokenizer & Snowball Stemmer)
Computer Vision & NLP
Advanced Research Portfolio
Certifications
Professional certifications in AI, Machine Learning, and specialized engineering mental tools.
Machine Learning Specialization
DeepLearning.AI • Sep 2023
A foundational 3-course specialization covering Supervised Learning, Advanced Learning Algorithms, and Unsupervised Learning techniques.
Included Courses
Learning How to Learn
Coursera • Aug 2023
Powerful mental tools to help master tough subjects.
Skills & Stack
A comprehensive overview of the tools and technologies I use to build and scale products.