Book a free call
Open for Q1 2027 engagements

Ship AI that runs in production.

I'm Kamil Ansari, an AI systems consultant and founder. I design, build and run AI that works under real load: voice agents that have made 1M+ sales calls, a multi-agent legal platform that turned 10 days of PhD-level patent analysis into 16 minutes, and real-time AI at Implied, a platform used by 5 of the 10 largest US hedge funds.

Book a free scoping call See the work
Consulting for Implied, San Francisco Voice AI showcased on Shark Tank India 4.99★ across 97 client reviews
Where my work has shipped
IImplied SAs seen on Shark Tank India NNewtuple MM3E Solutions TTellyFlow EExbida AAiVANTA
1M+
AI sales calls handled across finance, real estate and more
100+
Software projects shipped for clients in 10+ countries
400+
French hospitals running software I built
5/10
Largest US discretionary hedge funds use the platform I consult for (Implied)
Case studies

Systems in production. Numbers attached.

Production AI and software across legal tech, social care, healthcare, fintech, real estate and enterprise operations.

Multi-Agent Legal AI
Legal tech client · Built at Newtuple as Senior GenAI Engineer

Patent infringement & invalidation: 10 days of PhD-level analysis in 16 minutes

An AI platform that reads a patent the way an expert examiner does, then searches a 2TB corpus for prior art and infringing claims.

Challenge

Invalidity and infringement analysis needs specialists to compare each claim against millions of documents. A single matter tied up a team of ~400 PhD-level researchers for around 10 days.

Solution

I designed and built the multi-agent pipeline on heavy cloud infrastructure. It ingests and indexes 2TB of patent data, then uses specialised agents to assess novelty, obviousness, enablement and patent eligibility, with claim-by-claim citations. The design works with any model (Claude or GPT-4o) and is checked against a golden dataset written by attorneys.

Impact

Analysis that took a team of researchers 10 days now comes back in 16 minutes, so legal teams can review far more matters.

16 min
End-to-end analysis, down from ~10 days
2 TB
Patent data ingested & indexed
~400
PhD researchers' workload automated
4
Legal tests run per patent: novelty, obviousness, enablement, eligibility
AWSPythonLangGraphClaudeGPT-4oVector DBDistributed ingestionLLM evals
Voice AI · Founder
TellyFlow · Built for fintech, real estate, hospitality & HR

AI sales agent teams behind 1M+ calls, showcased on Shark Tank India

TellyFlow runs AI voice agents that call, qualify and close: cold outbound, site-visit booking, loan pre-screening and follow-ups, sounding local in 30+ languages.

Challenge

Cold calling converts at around 2%, and human call centres don't scale. Businesses needed agents that sound human, follow compliance scripts and hand hot leads to people at the right moment.

Solution

I founded TellyFlow and built the platform end to end. It has a campaign engine (upload a list, set the script and compliance rules), real-time voice agents with under 750ms latency, live analytics, and a smart handoff that passes transcripts and lead scores to the sales team.

Impact

1M+ (10 lakh+) calls handled across client campaigns, with 1,000+ running at once. A fintech client showcased TellyFlow's agent "Neha" to the Sharks on Shark Tank India as the backbone of its sales operation.

1M+
AI calls handled (10 lakh+)
1,000+
Calls running at the same time
64x
Client revenue vs. setup cost (₹3.22 Cr from ₹5L)
<750 ms
Voice response latency

"All our loan sales and financial advisory: powered by AI agents." A fintech client pitched on Shark Tank India Season 5, with TellyFlow's agent "Neha" as its sales engine.

TellyFlow client on Shark Tank India · S5 · 2026
PythonFastAPIRealtime STT/TTSLLM agentsTelephony / SIPDocker30+ languages
GenAI · High-Governance
Children's social care · United Kingdom · Built at Newtuple as Senior GenAI Engineer

AI referral management & automated reporting for children's care-home placements

An AI system that triages referrals for vulnerable children, matches each child to the best-fit care home and writes the reporting automatically, under strict UK data governance.

Challenge

Placement teams lost hours to manual triage: reading email referrals, scoring urgency and checking availability across homes. Reporting on every child and every home took more hours on top. The data is among the most sensitive there is.

Solution

I built the referral engine. It ingests referrals from email, uses AI to parse, triage and score each case against live placements, and recommends the best fit by needs, availability and matching criteria. I also automated GenAI reports for every child and every care home, with role-based access, audit trails and GDPR-grade controls throughout.

Impact

Less manual triage, better placement matching, and reporting that used to take days now runs automatically, deployed in a highly sensitive, heavily governed environment.

Email → match
Referrals parsed, triaged & scored against live placements
Auto-reports
GenAI reporting for every child & every home
RBAC · Audit trails · GDPR
Deployed under strict UK children's-services data governance
PythonFastAPILLM agentsEmail automationRBACAzure / AWSAudit logging
Healthcare Software
M3E Solutions · Paris, France

Excel-based clinical software rolled out across 400+ hospitals in France

M3E Solutions builds digital training and expert tools for healthcare professionals. They needed an Excel-based system that clinicians could use with no setup and no IT tickets.

Solution

I engineered a full software product inside Excel, with VBA automation, guided forms, validation and reporting, robust enough to roll out unchanged across hospital networks.

Impact

Now deployed in 400+ hospitals across France, with no installation for clinical staff.

400+
Hospitals in France
0
Installs needed: runs where clinicians already work
50+
Excel software products and automations built for clients worldwide. Read the M3E CEO's words in Client voices below.
Fintech AI · San Francisco · Ongoing

Consulting for Implied: real-time AI for public-markets investors

Consulting for a San Francisco AI research platform for public markets. I build AI-driven systems and live monitoring pipelines that process huge financial data streams (filings, live transcripts, market data) with low latency.

5 / 10largest US discretionary hedge funds use Implied
Legal AI · India · via Newtuple

Vidur: AI research agent for corporate, tax & regulatory law

Built for a legal-tech client: a research copilot that answers questions on Indian corporate, tax and regulatory law, with cited sources lawyers can trust.

Citedanswers from statutes & case law
Advanced RAG

AI GST legal bot with highlighted source PDFs

Answers GST and tax questions and opens the exact passage, highlighted, in the source law book. It includes an ETL ingestion pipeline, RBAC, admin panels, tracing and load-tested serverless infrastructure.

RBACadmin panels · LLMOps · load-tested
Agentic AI · MCP

AiVANTA: MCP server for autonomous AI ad creation

A Model Context Protocol server that lets AI agents plan, generate and launch ad creatives and campaigns end to end, with no manual design loop.

MCPagent-controlled ad pipeline
AI Platform

AI Bot Generator: any website into a chatbot API

Paste a URL and it crawls every page, file and image (with OCR), builds a vector store and returns a ready-to-use bot API in minutes.

URL → APIcrawl · OCR · vector store
Trading Tools · USA

Stock Keys: hotkey trading app for US traders

A multi-threaded Windows app that maps a trading platform's most-used actions to a dedicated 9-key keypad, shipped to traders across the USA.

9-keyhardware + software product

Most AI never leaves the notebook.

Production is where AI projects go to die. That's where I work.

Pilots stall at the demo. Mine run on phone lines with a thousand calls at once, inside UK children's-services governance, and across hospital networks in France.

I don't hand you a prototype and disappear. I design for scale, security and measurable ROI from day one, then stay until it's live and paying for itself.

See production work
Concurrency
1,000+ live calls
Voice latency
<750 ms response
Client revenue
64x the setup cost
Data scale
2TB patent corpus
What I do

Pick the problem. I'll ship the system.

Strategy, architecture and hands-on engineering from one senior builder. No hand-offs to juniors. Accountable from the first call to production.

AI Voice & Agentic Systems

Multiply your sales floor: agents that call, qualify and book meetings 24/7, then hand hot leads to your team with full context. Plus multi-agent workflows and MCP servers that give your AI real tools.

  • Outbound & inbound voice agents
  • Multi-agent orchestration (LangGraph)
  • MCP servers & tool-using agents
  • Human handoff, evals & tracing
Proven at 1,000+ concurrent calls on TellyFlow

Enterprise RAG & Document Intelligence

Turn terabytes of legal, financial or operational documents into cited answers your experts trust, with the governance enterprises require.

  • Advanced RAG with cited sources
  • Legal, patent & tax research agents
  • RBAC, audit trails & admin panels
  • Golden datasets & LLM evals
Proven on a 2TB patent corpus

AI Products & Custom Platforms

From zero to a live product: SaaS, ERPs, internal tools and admin panels, designed to scale and built to be handed over cleanly.

  • 0 → 1 MVPs in weeks
  • ERP, HR & payroll systems
  • FastAPI, React & cloud-native
  • AWS, Azure & GCP deployments
Founder of 3 products: TellyFlow, Exbida, SimplifyNote

Operations Automation

Remove the hours of manual work your team repeats every day, from hospital networks to pharmacies to trading desks.

  • Excel & VBA systems at hospital scale
  • Document & report automation
  • Desktop & RPA automation
  • Python & Apps Script pipelines
50+ Excel products built, one in 400+ hospitals
Client voices

97 reviews. 4.99★. Many came back for more.

Verified reviews from founders and operators in the US, UK, Germany, the Netherlands, France and Israel.

Kamil is really outstanding from all the others. He is very knowledgeable, gifted and always helpful. Over and above all his qualities, he just never gives up and finds a solution even for the most complicated and unforeseen issues. It is really a great pleasure to work with him.
HF
Hartmut Frankowski
CEO, M3E Solutions · Paris, France
How we work together

From idea to production in weeks, not quarters

01

Discover

Week 1

We map your workflow, data and goals, and find the AI use case with the clearest ROI.

02

Prototype

Weeks 2–4

A working prototype on your real data, with evaluation metrics agreed upfront.

03

Productionize

Weeks 4–10

Hardened, secured and deployed on your cloud, with RBAC, observability, tests and documentation.

04

Operate & scale

Ongoing

Monitoring, evals and continuous improvement, or a clean handover to your team.

Ways to work together

Start small. Scale when it works.

1 week · fixed fee

AI Production Audit

I review your AI idea or stalled pilot, your data and your stack. You get a go/no-go call, a target architecture and a costed build plan.

  • Feasibility & ROI assessment
  • Architecture & model selection
  • Costed roadmap you can act on
Request an audit
6–10 weeks · fixed scope

Build Sprint

From prototype to production on your cloud: evals, RBAC, observability and handover docs included. You own all the code.

  • Working prototype in 2–4 weeks
  • Production deployment on your cloud
  • Tests, docs & clean handover
Scope a sprint
Monthly retainer

Fractional AI Lead

Senior AI architecture plus hands-on engineering for teams without an in-house AI lead. Roadmap, delivery and hiring support.

  • Dedicated weekly capacity
  • Architecture & code reviews
  • Ship features, not slide decks
Discuss a retainer

LLMs & Agents

OpenAIClaudeGeminiLangChainLangGraphMCPLangfuse

Retrieval

Vector DBsHybrid searchOCRETL pipelinesEvals

Backend

PythonFastAPIJavaScriptSQLRESTReact

Cloud & Ops

AWSAzureGCPDockerLambdaLoad testing

Automation

Excel / VBAApps ScriptSeleniumRPA
Founder track

I build, sell and run my own products

I've had to sell what I build, ship it to real users and keep it running at scale. You get that founder judgment on your project, not just code.

Live

TellyFlow

AI agent teams for outbound sales

Voice agents that call, qualify and close across finance, real estate, hotels, tourism, HR and appointments. Showcased on Shark Tank India by a fintech client.

1M+ calls1,000+ concurrent30+ languages
tellyflow.com
Early access

Exbida

The ERP for UAE & GCC businesses

HR, payroll, finance and procurement on one employee record, one approval engine and one ledger, with the night audit and outlet controls hotels need.

3 modulesMulti-entityMulti-currency
exbida.com
Sunset

SimplifyNote

One-click note-taking for the web

A browser extension that sent highlights, quotes, screenshots and OCR text from web pages, PDFs and YouTube straight into Google Docs. It grew to 20K users before I moved on to AI.

20K usersOCR built inGoogle Cloud
About

Engineer by training. Founder by instinct.

I've spent six years shipping software for clients across the US, Europe, the UK, the Middle East and India: first as an independent consultant, then as a Senior GenAI Engineer leading enterprise AI delivery at Newtuple, and now as a founder and AI systems consultant.

Today I consult for Implied in San Francisco, whose AI research platform is used by 5 of the 10 largest US discretionary hedge funds, while running TellyFlow and Exbida.

My edge is that I've lived on both sides. I've built AI inside enterprise governance and built my own products from zero to thousands of users. You get senior engineering judgment and a founder's focus on outcomes.

B.Tech, Computer Science & EngineeringJamia Millia Islamia, New Delhi
Global rank 581 of 142,000+CodeChef July Long Challenge, data structures & algorithms
2026 – Present

Founder · Exbida

A complete HR, payroll, finance and procurement ERP for the UAE and GCC, now in pre-sales.

2025 – Present

Founder · TellyFlow

AI voice agent teams for outbound sales. 1M+ calls, showcased on Shark Tank India by a fintech client, campaigns live in finance, real estate and hospitality.

May 2025 – Present

AI Systems Consultant · Implied, San Francisco

Consulting on AI-driven systems and low-latency monitoring pipelines for Implied's AI research platform for public-markets investors, used by 5 of the 10 largest US discretionary hedge funds. I own features end to end, from build and testing to release and production support.

May 2024 – May 2025

Senior GenAI Engineer · Newtuple Technologies, Pune

Led delivery of enterprise GenAI systems: patent infringement and invalidation (2TB, 10 days → 16 min), UK children's placement referrals, Vidur, a GST legal bot, advanced RAG, RBAC and admin platforms. Worked across LLMOps, agentic frameworks, voice bots and scalable cloud infrastructure.

2022 – 2023

Founder · SimplifyNote

A one-click note-taking extension for web, PDF and YouTube, deployed on Google Cloud. Grew to 20K users.

2020 – Present

Independent Software Consultant · Global clients

100+ projects delivered worldwide, from feasibility studies to finished products: 50+ Excel software products, Windows automation, AI bots and LLM apps. Fiverr Level Two, 4.99★ across 97 reviews, with many repeat clients.

FAQ

What clients ask before the first call

You work directly with the person who architects and writes the system. No account managers, no hand-offs to juniors. I've shipped AI at enterprise scale and built my own products, so I care about the business result, not just the ticket.

Typically a working prototype on your data in 2–4 weeks and a production deployment in 6–10 weeks. Smaller automations and Excel products often ship within days.

Yes. I've deployed AI on UK children's social-care data and on confidential legal corpora. Role-based access, audit trails, private cloud or self-hosted models, and GDPR-aligned practices are standard in what I build.

Legal tech, financial services and fintech, healthcare and social care, real estate, hospitality, HR and recruiting, and e-commerce and ads. The patterns (agents, retrieval, automation) carry across domains.

Three ways: a one-week AI Production Audit, a fixed-scope Build Sprint (6–10 weeks), or a monthly Fractional AI Lead retainer. Every engagement starts with a free 30-minute scoping call, and you own all the code.

I'm based in India (IST) and work regularly with teams in the US, UK, Europe and the Middle East, with overlapping hours and async updates.

Let's build

Have an AI project stuck at demo stage? Let's ship it.

Book a free 30-minute scoping call. You'll leave with a straight answer on feasibility, a rough architecture, a timeline and a budget range, whether or not we work together.

Book a free scoping call Message on WhatsApp
Email
ansarikamil086@gmail.com Phone / WhatsApp
+91 80762 21529 Fiverr
Level Two · 4.99★ · 97 reviews