Full-stack
My goal is to write maintainable, clean and understandable code to make the development process enjoyable.
My goal is to write maintainable, clean and understandable code to make the development process enjoyable.
Hello! I'm Asish, a full-stack developer and Founding Engineer.
With 4+ years of experience building AI-integrated systems, video-first platforms, and distributed systems at scale. I specialize in transforming product visions into high-performance applications with cutting-edge technology.
OpenAI / Olamma / CrewAI / AutoGen / LangGraph / LangChain / PyTorch
NodeJs / Flask / GraphQL / REST API / Docker / Heroku
AWS / Azure / IaC / Elastic Load Balancer/ PostgreSQL / MySQL / AWS RDS / CosmosDB / Supabase / Firebase
These are Some of my favourite technologies, topics or tools I've worked on;
Android SDK / Kotlin / Java / Flutter / XML / JSON / Crashlytics / Google Console / GoogleAPI's
Founding Engineer | Python(Flask), AWS, PostgreSQL, LLMs, Agentic AI, ML, Redis, Docker
AI Engineer | ReactJS, Flask, LangGraph, CrewAI, OpenAI API, Pinecone, Supabase
Graduate Research Assistant | Node.JS, Docker
Software Development Engineer | Java, Kotlin, Google APIs, GraphQL, Apollo, Firebase, Crashlytics
Backend Engineer | Node.JS, MySQL, AWS, Docker, Microservices
Frontend developer | React.JS, BootStrap, PostgreSQL, RDP Protocol
Work experience
4+ years
Led end-to-end development of Kaana from POC to production on iOS & Android Platforms for a regional user base. Architected & deployed infrastructure supporting 10K+ users, including auth, builds, API endpoint, EC2 deployment, PostgreSQL DB, Database design, Redis caching, CI/CD pipelines, docker, and Posthog analytics.
Engineered a low-latency LLM inference pipeline using Gemini API and custom ML ranking algorithms, processing natural language queries into optimized SQL filters with <500ms response time. Architected an adaptive ML-driven search system for restaurants, reels, and photos, achieving 95%+ relevance.
Built an AI-BDR platform that can automate outreach campaigns for businesses, converting 7 leads into paying users. Orchestrated LangGraph agents to research, generate & send them to Gmail drafts for each target saving 70% time.
Integrated advanced AI workflows for prospect research and personalized campaign generation. Implemented comprehensive automation system for business development representatives.
Built an AI-stock Advisor system that talks to users about their portfolios by fine tuning a Llama model on SEC filings and news articles, with RAG & Langchain agents for real-time stock data, deployed it on Azure Containers.
Integrated advanced AI capabilities for portfolio analysis and investment recommendations. Implemented comprehensive financial advisory system with natural language processing for user interactions.
Launched Google Maps features to users to pin the delivery location & provide live tracking of delivery partners. Integrated Google APIs to build an automated route scheduling system for drivers increasing delivery rate by 30%.
Optimized 13 APIs to GraphQL, reducing latency by 60%, improving user experience for 30 million+ customers. Implemented data caching & GraphQL queries for efficient data retrieval.
Built a distributed database in C++ with serializable snapshot isolation, utilizing Multi-Version Read Consistency. Implemented fault-tolerant replication using Available Copies algorithm and lock-free transaction commits.
Advanced distributed systems project with sophisticated concurrency control mechanisms for handling concurrent transactions across multiple sites.
Published a research paper assessing 6 major LLMs on their code generation capabilities, comparing results runtime, memory, time and space complexity assessing their performance based on structured LeetCode problems ranging from easy to hard.
Built an academic portal for students to submit their assignments to generate keyword graphs helping them assess their assignment comparing it to what professor is looking for. It is also useful for professors in reducing grading time evaluating keyword graphs.
Engineered a machine learning model to predict NBA game outcomes on 26K+ games to achieve 94.5% accuracy. Implemented advanced feature engineering and ensemble methods to analyze player statistics, team performance, and historical data.
Built comprehensive data pipeline for processing NBA statistics and game data. Developed sophisticated prediction system with high accuracy for game outcome forecasting.
Handcrafted by Myself
Simple Website built on HTML/ CSS & JS
Hosted on Github
Designed by Taisia /
