[{"data":1,"prerenderedAt":395},["ShallowReactive",2],{"index":3,"projects":116},{"id":4,"title":5,"body":6,"description":7,"experience":88,"extension":94,"hero":95,"intro":97,"meta":98,"navigation":28,"path":99,"seo":100,"services":101,"stats":107,"stem":112,"work":113,"works":114,"__hash__":115},"index/index.yml","Richie Patil – Developer & Creative Solutions Architect",{"title":5,"description":7,"seo":8,"hero":9,"intro":17,"stats":22,"work":39,"services":43,"experience":63,"works":86},"Developer and Creative Solutions Architect based in Mumbai, specialising in Python-based ML, Generative AI, MLOps, and scalable backend architectures.",{"title":5,"description":7},{"role":10,"location":11,"status":12,"lead":13,"stack":14},"Developer & Creative Solutions Architect","Mumbai","Available for work","I build robust software systems and intelligent ML — scalable APIs and data pipelines that turn raw data into decisions.",[15,16],"Python · TypeScript · Node.js · React","AWS · Docker · MongoDB · Git",{"lead":18,"body":19,"quote":20,"beyond":21},"Hey, I’m Richie — a developer and creative solutions architect in Mumbai. I build APIs that feel effortless and models that earn their place in production.","Node.js and TypeScript on the backend, Python for machine learning and generative AI, AWS and Azure underneath. Everything from high-volume supply-chain SaaS to enterprise insurance systems.","Start with the why. Then make it fast.","Beyond the screen: strategy games, running Mumbai’s routes, and home cooking.",[23,29,32,35],{"value":24,"decimals":25,"suffix":26,"caption":27,"accent":28},99.7,1,"%","Reduction in API response time",true,{"value":25,"suffix":30,"caption":31},"M+","Certificate capacity, Sankalp platform",{"value":33,"caption":34},5,"SaaS products built on my APIs",{"value":36,"suffix":37,"caption":38},10,"K+","Products tracked daily by my data pipeline",{"heading":40,"note":41,"span":42},"Building systems, solving problems.","❋ 72Street, TavcoGrowth and Sankalp are live; the rest are private, academic or archived.","2022 — 2026",{"heading":44,"body":45,"items":46},"Systems that scale, and models that tell you something.","I work where backend engineering meets machine learning. Most projects draw on more than one of these.",[47,51,55,59],{"title":48,"body":49,"image":50},"Backend & APIs","Node.js and TypeScript services built for load. Clean contracts, fast responses, thorough documentation.","/content/projects/sankalp.png",{"title":52,"body":53,"image":54},"Machine learning","Predictive models with scikit-learn, Pandas and NumPy, from exploratory analysis to production.","/content/projects/sunspots.jpg",{"title":56,"body":57,"image":58},"Generative AI & RAG","LLM systems with LangChain and the OpenAI API that answer from your own data.","/content/projects/rag.webp",{"title":60,"body":61,"image":62},"Data & MLOps","ETL, aggregation pipelines, Docker, CI/CD and serverless on AWS and Azure.","/content/projects/Flutter.webp",{"note":64,"education":65,"items":66},"Supply-chain SaaS, enterprise insurance, data science.","Before that: B.Tech in Computer Science (AI & ML), Amity University Mumbai, 2020–2024.",[67,72,76,81],{"org":68,"role":69,"when":70,"note":71},"Sunday Tech","Backend Developer","Sep 2024 — Present","APIs for 5 SaaS apps. Cut latency 99.7%, from 45s to 145ms. Led a team of 3 across two projects.",{"org":68,"role":73,"when":74,"note":75},"Backend Developer Intern","Jun — Sep 2024","Resolved 20+ production issues, 17% faster. Documented 25+ third-party integrations.",{"org":77,"role":78,"when":79,"note":80},"WNS Global Services","Intern Trainee (WTS)","Jan — May 2024","Enterprise insurance testing: 1,000+ issues logged, 450+ test scenarios.",{"org":82,"role":83,"when":84,"note":85},"Celebal Technologies","Data Science Intern","Aug — Dec 2023","EDA on 500K+ records behind two marketing strategies. ETL on 10GB+ of data.",{"heading":40,"body":87},"Scalable backends and intelligent ML, from concept to deployment, plus a few experiments along the way.",{"note":64,"education":65,"items":89},[90,91,92,93],{"org":68,"role":69,"when":70,"note":71},{"org":68,"role":73,"when":74,"note":75},{"org":77,"role":78,"when":79,"note":80},{"org":82,"role":83,"when":84,"note":85},"yml",{"role":10,"location":11,"status":12,"lead":13,"stack":96},[15,16],{"lead":18,"body":19,"quote":20,"beyond":21},{},"/",{"title":5,"description":7},{"heading":44,"body":45,"items":102},[103,104,105,106],{"title":48,"body":49,"image":50},{"title":52,"body":53,"image":54},{"title":56,"body":57,"image":58},{"title":60,"body":61,"image":62},[108,109,110,111],{"value":24,"decimals":25,"suffix":26,"caption":27,"accent":28},{"value":25,"suffix":30,"caption":31},{"value":33,"caption":34},{"value":36,"suffix":37,"caption":38},"index",{"heading":40,"note":41,"span":42},{"heading":40,"body":87},"ptqz_kREh0jKxJkdwyx8CbpfM6BnfMethUlJAnOhHm4",[117,168,210,233,277,301,324,347,371],{"id":118,"title":119,"build":120,"confidential":121,"extension":94,"gallery":122,"image":138,"imagePortrait":139,"link":140,"linkLabel":141,"meta":142,"metric":157,"metricLabel":158,"order":25,"outcome":156,"problem":155,"slug":145,"stem":165,"summary":154,"tags":166,"type":146,"year":147,"__hash__":167},"projects/projects/7.yml","72Street.ai","A phone and tablet app with OTP login, advisor recommendations that carry an entry range, target, stop-loss and risk-reward ratio, live market data and watchlists, and subscription plans. Consent-first and DPDP-compliant, with no dark patterns.",false,[123,126,129,132,135],{"src":124,"alt":125},"/content/projects/72street/phones.webp","72 Street profile, market data and subscription screens on phones",{"src":127,"alt":128},"/content/projects/72street/tablet-1.webp","Market data and watchlist on tablet",{"src":130,"alt":131},"/content/projects/72street/tablet-2.webp","Advice list on tablet",{"src":133,"alt":134},"/content/projects/72street/tablet-3.webp","Advice cards with potential left and subscribe actions on tablet",{"src":136,"alt":137},"/content/projects/72street/tablet-4.webp","Profile and subscription details on tablet","/content/projects/72street/cover.webp","/content/projects/portrait/72street.webp","https://play.google.com/store/apps/details?id=com.everisseai.app72street","View on Google Play ↗",{"path":143,"body":144},"/projects/7",{"slug":145,"order":25,"title":119,"type":146,"year":147,"image":138,"imagePortrait":139,"link":140,"linkLabel":141,"tags":148,"summary":154,"problem":155,"build":120,"outcome":156,"metric":157,"metricLabel":158,"gallery":159},"72street","Fintech platform","2026",[149,150,151,152,153],"Fintech","Mobile & tablet","Market data","Subscriptions","SEBI-aligned","An advisory-only stock app for Indian investors: human-guided recommendations, with the reasoning behind every call.","Most investing apps are built for speed and execution. 72 Street wanted the opposite: calm, explainable, advisor-led guidance, with no trading, demat or order execution in the loop.","Live on Google Play, rated 4.9 by its early users, with 500+ downloads.","4.9★","Google Play rating",[160,161,162,163,164],{"src":124,"alt":125},{"src":127,"alt":128},{"src":130,"alt":131},{"src":133,"alt":134},{"src":136,"alt":137},"projects/7",[149,150,151,152,153],"H1JXUDJ0oSxz2Sxhf59uL4Jg1JOlgvpLDyeESTEe_k4",{"id":169,"title":170,"build":171,"confidential":121,"extension":94,"gallery":172,"image":182,"imagePortrait":183,"link":184,"linkLabel":185,"meta":186,"metric":201,"metricLabel":202,"order":190,"outcome":200,"problem":199,"slug":189,"stem":207,"summary":198,"tags":208,"type":191,"year":147,"__hash__":209},"projects/projects/8.yml","TavcoGrowth","A multi-tenant CRM where a caller’s whole day is one screen: callbacks first, tap to call, log the outcome, move on. Each company’s data is kept separate at the database level, so isolation can’t be switched off by mistake, and every enquiry from the landing pages flows straight in.",[173,176,179],{"src":174,"alt":175},"/content/projects/tavco/02.webp","GrowthPipe product section: a caller's whole day is one screen",{"src":177,"alt":178},"/content/projects/tavco/03.webp","Services: Meta lead generation, Google Ads, WhatsApp automation, websites",{"src":180,"alt":181},"/content/projects/tavco/04.webp","Team section with engineering and marketing figures","/content/projects/tavco/01.webp","/content/projects/portrait/tavco.webp","https://tavcogrowth.com/",null,{"path":187,"body":188},"/projects/8",{"slug":189,"order":190,"title":170,"type":191,"year":147,"image":182,"imagePortrait":183,"link":184,"tags":192,"summary":198,"problem":199,"build":171,"outcome":200,"metric":201,"metricLabel":202,"gallery":203},"tavcogrowth",2,"SaaS CRM",[193,194,195,196,197],"Node.js","TypeScript","Express","Sequelize","Multi-tenant SaaS","GrowthPipe, a lead-management CRM, and the site that feeds it: one partner for the software and the leads.","Leads arrived from Meta and Google ads, WhatsApp and landing pages, and callers lost track of who to call back and what had already happened.","Tavco’s own team ran on it every day before it was sold to anyone, and it is now the product at the centre of Tavco Growth.","Multi-tenant","Isolation built into the database",[204,205,206],{"src":174,"alt":175},{"src":177,"alt":178},{"src":180,"alt":181},"projects/8",[193,194,195,196,197],"FjR-y0vO86XKjLcpQwNi7AUK_8U8OfEAFta7y2kVLq0",{"id":211,"title":212,"build":213,"confidential":28,"extension":94,"gallery":185,"image":185,"imagePortrait":185,"link":185,"linkLabel":185,"meta":214,"metric":228,"metricLabel":229,"order":218,"outcome":227,"problem":226,"slug":217,"stem":230,"summary":225,"tags":231,"type":219,"year":147,"__hash__":232},"projects/projects/9.yml","LADM NCL Portal","Separate asset and land survey flows, with online, offline and manual-entry records in one place. Long, structured forms capture each survey, and searchable tables track every record from compensation initiated to approval by the nodal officer.",{"path":215,"body":216},"/projects/9",{"slug":217,"order":218,"title":212,"type":219,"year":147,"confidential":28,"tags":220,"summary":225,"problem":226,"build":213,"outcome":227,"metric":228,"metricLabel":229},"ladm-ncl",3,"Enterprise portal",[221,222,223,224],"Enterprise","Forms","Data management","Workflows","A land-acquisition and resettlement portal built on extensive form filling, data manipulation and record management.","Asset and land surveys for resettlement and rehabilitation produce large, detailed records that pass through several hands before compensation can be approved.","One portal for the whole survey-to-compensation workflow, in place of scattered spreadsheets and paper.","2 surveys","Asset and land, in one portal","projects/9",[221,222,223,224],"D0VeCA8RWBim23SFzlOGFuDku8EaI1Rubkn6hjzynPg",{"id":234,"title":235,"build":236,"confidential":121,"extension":94,"gallery":237,"image":250,"imagePortrait":251,"link":252,"linkLabel":185,"meta":253,"metric":267,"metricLabel":268,"order":257,"outcome":266,"problem":265,"slug":256,"stem":274,"summary":264,"tags":275,"type":258,"year":259,"__hash__":276},"projects/projects/1.yml","Sankalp NGO Platform","React and TypeScript on the front, Node.js and SQL behind it, on a Linux VPS. Authentication encrypts and decrypts sensitive data end to end, and DevSecOps practice is built into every deploy.",[238,241,244,247],{"src":239,"alt":240},"/content/projects/sankalp/02.webp","Sankalp projects section: Drishti Mitra, Nari tu Narayani, Dhara and Sharda",{"src":242,"alt":243},"/content/projects/sankalp/03.webp","Stories from the field, written by volunteers",{"src":245,"alt":246},"/content/projects/sankalp/04.webp","Impact counters: drives conducted, education drives, slums adopted, government cards made",{"src":248,"alt":249},"/content/projects/sankalp/05.webp","More impact counters: health camps, cleanup drives, saplings planted, members active","/content/projects/sankalp/01.webp","/content/projects/portrait/sankalp.webp","https://thesankalp.org",{"path":254,"body":255},"/projects/1",{"slug":256,"order":257,"title":235,"type":258,"year":259,"image":250,"imagePortrait":251,"link":252,"tags":260,"summary":264,"problem":265,"build":236,"outcome":266,"metric":267,"metricLabel":268,"gallery":269},"sankalp",4,"Full-stack platform","2025",[261,194,193,262,263],"React","SQL","DevSecOps","A high-performance full-stack platform for managing certificates and user data at scale.","The Sankalp Organisation needed one reliable system to issue, store and verify certificates and user records, at a volume manual processes couldn’t keep up with.","Headroom for over a million certificates, and a live site the team runs every day.","1M+","Certificate capacity",[270,271,272,273],{"src":239,"alt":240},{"src":242,"alt":243},{"src":245,"alt":246},{"src":248,"alt":249},"projects/1",[261,194,193,262,263],"5SR2wDKvEmFzqOSUJWiVPHE4CEX6OBQWWzkj_nEmHMI",{"id":278,"title":279,"build":280,"confidential":121,"extension":94,"gallery":185,"image":58,"imagePortrait":281,"link":185,"linkLabel":185,"meta":282,"metric":296,"metricLabel":297,"order":33,"outcome":295,"problem":294,"slug":285,"stem":298,"summary":293,"tags":299,"type":286,"year":259,"__hash__":300},"projects/projects/2.yml","RAG Decision Assistant","Documents are embedded for vector search. LangChain orchestrates retrieval and generation with Llama 3 or GPT. It runs serverless on AWS Lambda behind API Gateway, scales automatically, and sits behind enterprise-grade authentication.","/content/projects/portrait/rag.webp",{"path":283,"body":284},"/projects/2",{"slug":285,"order":33,"title":279,"type":286,"year":259,"image":58,"imagePortrait":281,"tags":287,"summary":293,"problem":294,"build":280,"outcome":295,"metric":296,"metricLabel":297},"rag","AI system",[288,289,290,291,292],"Python","LangChain","AWS Lambda","RAG","Llama 3 / GPT","A retrieval-augmented system for querying risk documents in real time and supporting automated decisions.","Risk decisions depended on dense documents that people had to read and cross-check by hand.","Answers grounded in the source documents, available on demand instead of after a manual review.","Serverless","Auto-scaling on AWS Lambda","projects/2",[288,289,290,291,292],"aLP0PRYOj6mxtXZItV1o-NsNen2np05ruknpIECqb2U",{"id":302,"title":303,"build":304,"confidential":121,"extension":94,"gallery":185,"image":54,"imagePortrait":305,"link":185,"linkLabel":185,"meta":306,"metric":319,"metricLabel":320,"order":310,"outcome":318,"problem":317,"slug":309,"stem":321,"summary":316,"tags":322,"type":311,"year":312,"__hash__":323},"projects/projects/3.yml","Sunspot & Solar Weather Prediction","Feature engineering on observatory and satellite records, then time-series and ensemble models compared and tuned against each other.","/content/projects/portrait/sunspots.webp",{"path":307,"body":308},"/projects/3",{"slug":309,"order":310,"title":303,"type":311,"year":312,"image":54,"imagePortrait":305,"tags":313,"summary":316,"problem":317,"build":304,"outcome":318,"metric":319,"metricLabel":320},"sunspots",6,"ML research","2024",[288,52,314,315],"Time series","Research","My final-year dissertation: forecasting sunspot activity and space weather with machine learning.","Solar activity disrupts satellites, communications and power grids, and it is notoriously hard to forecast.","Around 85% prediction accuracy, presented at the university symposium.","~85%","Prediction accuracy","projects/3",[288,52,314,315],"hUMd8WCVjoSI2FX4igeupmBkI4IplkJgY-Ll-Mc0JUM",{"id":325,"title":326,"build":327,"confidential":121,"extension":94,"gallery":185,"image":185,"imagePortrait":185,"link":185,"linkLabel":185,"meta":328,"metric":342,"metricLabel":343,"order":332,"outcome":341,"problem":340,"slug":331,"stem":344,"summary":339,"tags":345,"type":333,"year":312,"__hash__":346},"projects/projects/4.yml","E-Commerce Web Scraper","Selenium renders the dynamic pages and BeautifulSoup parses them. Rate limiting and incremental updates keep the pipeline polite and efficient, and the cleaned data flows into Power BI.",{"path":329,"body":330},"/projects/4",{"slug":331,"order":332,"title":326,"type":333,"year":312,"tags":334,"summary":339,"problem":340,"build":327,"outcome":341,"metric":342,"metricLabel":343},"scraper",7,"Data pipeline",[288,335,336,337,338],"Selenium","BeautifulSoup","ETL","Power BI","A scalable scraper and ETL pipeline feeding an e-commerce analytics dashboard in Power BI.","Product and pricing data sat behind dynamic pages that simple scrapers couldn’t read.","More than 10,000 products tracked daily, with no manual collection.","10K+","Products tracked daily","projects/4",[288,335,336,337,338],"fqRbjIPN5TfEU5lzn73bUmyNrziUiK1uskIFhXUM_7A",{"id":348,"title":349,"build":350,"confidential":121,"extension":94,"gallery":185,"image":62,"imagePortrait":185,"link":185,"linkLabel":185,"meta":351,"metric":366,"metricLabel":367,"order":355,"outcome":365,"problem":364,"slug":354,"stem":368,"summary":363,"tags":369,"type":356,"year":357,"__hash__":370},"projects/projects/5.yml","Amizone App","I reverse-engineered the portal’s APIs, then built a cross-platform Flutter client with secure login and offline caching.",{"path":352,"body":353},"/projects/5",{"slug":354,"order":355,"title":349,"type":356,"year":357,"image":62,"tags":358,"summary":363,"problem":364,"build":350,"outcome":365,"metric":366,"metricLabel":367},"amizone",8,"Mobile app","2023",[359,360,361,362],"Flutter","Dart","Mobile","API integration","A Flutter app recreating Amity’s student portal, built on reverse-engineered APIs.","The legacy portal was slow and awkward on phones, and phones are where students actually use it.","A mobile-first redesign of a legacy portal on iOS and Android, from one codebase.","1 → 2","One codebase, two platforms","projects/5",[359,360,361,362],"ZjKYA9KesZAcZ_lqgi99bdN0c4cQmmmfnswAKMlzmdg",{"id":372,"title":373,"build":374,"confidential":121,"extension":94,"gallery":185,"image":375,"imagePortrait":185,"link":185,"linkLabel":185,"meta":376,"metric":390,"metricLabel":391,"order":380,"outcome":389,"problem":388,"slug":379,"stem":392,"summary":387,"tags":393,"type":381,"year":382,"__hash__":394},"projects/projects/6.yml","Journey of the Ring","I storyboarded every branch, then built custom art and UI in Ren’Py.","/content/projects/jotr.jpeg",{"path":377,"body":378},"/projects/6",{"slug":379,"order":380,"title":373,"type":381,"year":382,"image":375,"tags":383,"summary":387,"problem":388,"build":374,"outcome":389,"metric":390,"metricLabel":391},"jotr",9,"Game","2022",[384,288,385,386],"Ren’Py","Game dev","Interactive fiction","A Lord of the Rings parody visual novel with branching endings.","A creative side project, and a lesson in pacing, player choice and story structure.","A playable story with multiple endings, and a reminder that range beyond the backend matters.","Branching","Multiple endings","projects/6",[384,288,385,386],"4WcZTDeUPD867MYZM0YcnMX_9n27LW9f3bE-Ox3q9qs",1790582864391]