Hire the Best Artificial Intelligence Engineers in Pune, IN

You get a PM and a builder in one. I plan consumer AI products and I prototype them myself. Working flows, not slide decks. Consumer AI is a design problem first. Here is what shipping AI for consumers taught me. The model is rarely the bottleneck. The experience is. Most teams bolt a chat box onto their app and call it AI. Users do not want a chat box. They want the app to already know. So I design AI around constraint, not open-endedness. Bounded inputs. Predictable outputs. The less the AI improvises, the faster the product earns trust. And trust is the whole game in consumer AI. You lose it in one bad answer. You win it back over weeks, if ever. The best AI design is quiet. The user should feel smart, not the model. That belief changes everything downstream. Onboarding. Empty states. Latency handling. How you phrase a recommendation. Retention lives in those details, so I sweat them. Design is my second seat, not a side interest. I lead UX alongside product at my current company. Onboarding flows. First-session experience. The rituals that turn a download into a habit. Community dynamics and the privacy trade-offs that come with them. I prototype daily in clickable HTML using AI tools, so teams react to something real instead of debating abstractions. Growth is where the truth comes out. I run experimentation the boring, disciplined way. One steering metric at a time. Feature flags and phased rollouts in PostHog. Clean cohorts. Funnels that show you exactly where day one dies. Ship small. Measure honestly. Kill what does not move the number. Most consumer growth problems are activation problems wearing a costume. I start there. The AI plumbing, I do myself. Prompts. Evals. Guardrails. Latency. I write them, test them, and fix them. So when your AI feature misbehaves, I can tell you whether it is a model problem, a data problem, or an architecture problem. In my experience, it is usually architecture. The parts you cannot put in a skills tag. Ownership looks like this. Our repo once leaked credentials on a weekend night. I did not wait for morning or for permission. I ran the response, coordinated the fix, and wrote the postmortem before the workweek started. Commitment looks like this. When my current company hit a hard stretch and cut hours across the team, I did not quietly start shopping around. I told the founder I was staying through it. I am still there, still shipping. Quality looks like this. We have no dedicated QA team. So I review releases myself, hunt for the missing scenarios, and challenge my own requirements before engineers ever see them. If it ships broken, that is on me. I like it that way. Communication looks like this. My team and my stakeholders sit ten hours apart. Everything I write has to work while I sleep. So I write short, clear, and complete. My engineers actually read my docs. Ask them. Learning looks like this. When our AI pipeline kept failing, I did not shrug and blame the model. I documented the distinct failure patterns and turned each one into a fix. Mistakes become documentation, or they repeat. Background: AI PM on an AI scriptwriting tool (0 to 1). Founding PM at an e-commerce SaaS. Growth strategy at Merkle (Dentsu). Years of async India to US collaboration, so timezone overlap is a solved problem. Where I fit: 0 to 1 consumer AI products that need a first PM AI features that ship but do not stick Rescue work on AI that is slow, unreliable, or hallucinating Teams with no designer yet who need product and UX in one seat Not a fit: pure people management, enterprise B2B sales support, or a roadmap deck nobody will build. Tell me what you are building and where it is stuck.

Most AI projects die between the demo and production. I close that gap: LLM agents, RAG, text-to-SQL, and the data pipelines and warehouses underneath them. 70 contracts, 1000+ hours, Top Rated. Working in data since 2012. MTech in Data Science from BITS Pilani. 8 years at Accenture, 4 at Persistent Systems. Most AI projects stall in the same place. The demo works, then it meets real data. That gap is where I spend my time: ingestion, modelling, evaluation, and the unglamorous reliability work that decides whether an agent survives contact with production. WHAT I BUILD AI and LLM systems - LLM agents and multi-agent workflows, with tool use, retrieval and evaluation harnesses - Text-to-SQL over real warehouses, with guardrails so business users can self-serve - RAG pipelines: chunking, embeddings, retrieval quality, hallucination control - AI features embedded into existing products, not standalone demos Data engineering - ETL and ELT pipelines from source systems into the warehouse - Data quality, validation and pipeline observability - Warehouse and lakehouse modelling - Platform migrations and modernization Analytics people actually use - Apache Superset and Preset, across multiple long-running engagements including 100+ hour front-end builds - Executive dashboards, embedded analytics, custom visualizations - Reporting layers that keep working after handover RECENT WORK - Text-to-SQL interface that lets executives query the warehouse in plain English - ESG data ingestion and automation for compliance reporting - Predictive maintenance models for manufacturing - Superset front-end and visualization work across several repeat clients STACK Python, SQL, Claude and OpenAI APIs, LangChain, FastAPI, PostgreSQL, MySQL, Apache Superset, Apache NiFi, Airflow, Elasticsearch, Power BI, AWS, Docker HOW I WORK Small scope first. I would rather prove the approach on a paid discovery or one narrow slice than write a long proposal for something neither of us has tested. You get working code, documentation and a real handover, not a dependency on me. Good fit if you have real data, a real business problem, and you want something running rather than a prototype. Not a fit if you need someone to fill a seat and work a ticket queue. Tell me what you are building and where it is stuck. I will tell you honestly whether I am the right person for it.

A flashy AI demo and a system your business can trust are two very different things. The gap between them is engineering discipline, and that's where I work. I build production-grade AI for businesses, and can take it from idea to shipped product: RAG systems that let you use your own knowledge, AI agents that automate real workflows, and the web, mobile, and backend around them, so you get a working product, not just a model. The part most "AI experts" skip is the part I obsess over: clean data, reliable retrieval, real observability, sensible guardrails, and a human in the loop where it counts. Demos are easy. Systems that hold up under live traffic are not. Five years building production AI across Tokyo and India, most recently on Korn Ferry's Digital team, where I cut response latency ~40% and lifted retrieval accuracy 50%+ on live products, and earlier shipped computer vision systems at 80%+ accuracy. I work fluently in English and Japanese, so I bridge teams and markets, not just code. What I can build for you: • RAG and Graph RAG systems (re-ranking, metadata filtering, hybrid search) • AI agents and multi-agent workflows (MCP, tool use) • Computer vision (object detection, segmentation, OCR) • Full delivery: APIs, cloud, web and mobile, plus the production layer, deployment, observability, evaluation, and cost control Tell me what you're trying to build and I'll give you an honest read on whether it's worth it and how I'd approach it. I'd rather build something that lasts than ship a demo.

I work with businesses that need to build, evolve, or stabilize software systems. This could be a new product, a set of features, or an existing platform that is not performing the way it should. I lead Metisent, an AI-native software engineering agency focused on workflow automation, AI integration, backend systems, and custom platform development. Most client engagements run through Metisent, where we bring a small, focused team with senior involvement at every stage. Over the last 20+ years, I have worked hands-on and in leadership roles across financial services, healthcare, e-commerce, and martech. I co-founded Coditas, where I helped scale the engineering organization from 10 to 900+ people. Typical work we get involved in: - Backend and API development - AI-based applications and workflow automation - MVP builds and early-stage products - Improving or stabilizing existing systems - Debugging and resolving complex issues What clients generally find useful is not just execution, but the ability to: - Quickly understand what's actually going wrong or missing - Make practical architectural decisions early - Keep solutions simple and aligned with actual business needs If you have something that needs to be built, improved, or fixed, happy to take a look. All project engagements are handled through Metisent.

AI/ML Engineering & Generative AI Specialist | Production Deployment (AWS/Docker) As a proven AI/ML Engineer, I specialize in building, deploying, and scaling cutting-edge machine learning solutions that directly impact business revenue and user engagement. I turn complex prototypes into robust, production-ready systems. Core Expertise: Generative AI: LLMs (GPT, Gemini), RAG, Fine-Tuning, Virtual Try-On (VITON), and Voice Cloning for personalized experiences. Computer Vision & Search: Implementing high-impact features like visual search (Fashion-CLIP/OpenSearch) and deep image analysis. Data Optimization: Developing AI-driven dynamic pricing models leveraging real-time data to optimize revenue and inventory. MLOps & Cloud: Full-stack deployment using Python, Docker, AWS (S3, EC2), OpenSearch, and containerized architectures. Results Focused: Experience includes engineering visual search features, deploying a multilingual speech pipeline, and automating back-office operations (reducing manual time by 20+ hours/week). I deliver reliable, optimized AI solutions designed for measurable business impact. Let's discuss your project.

I’m a Computer Engineer by degree and a writer by passion, and I've tackled everything from whitepapers and blogs to academic theses and business proposals. If it needs to be written (especially on cybersecurity), I’m your go-to! I’m all about staying organized, diving deep into research, and hitting deadlines like they’re going out of style. Expect your projects done with time to spare—giving you plenty of room for feedback. I take real pride in delivering polished, top-quality work. Let’s get things done!