AI & Gen AI Engineering · Bangalore, India · 9+ Years

We Build the
Intelligence
Behind the Idea.

From concept to production — custom AI models, agentic systems, and Gen AI products built for startups, enterprises, and every bold idea in between. Not just a vendor. A partner.

85%
Junk CVs filtered
90%
Match accuracy
35%
Faster hiring
40%
Less interviewer load
9+
Years AI / Gen AI
85%
Noise filtered · Kotak
100%
Happy Clients
100%
In-house AI
🧠
90%
Match Accuracy
35%
Faster Hiring
🚀
9+ yrs
AI Experience
INPUT HIDDEN HIDDEN OUTPUT CV JD DB HR RANK MATCH
Data Ingestion
Model Training
Deployment
Trusted by
Kotak Mahindra Bank
TQCert · Tata Projects
TPLSPL · Tata Projects
Stealth EdTech Startup
What We Do

Full-Stack AI
Engineering

We don't just consult — we build. From in-house model training and data annotation to production-ready agentic systems. Every capability, owned end-to-end.

agent_pipeline.py
1from curioso import AgenticAI
2
3# Build your custom agent
4agent = AgenticAI(
5  model="custom-ft-v2",
6  data=your_domain_data,
7  deploy=True
8)
9agent.run()
🧠
Custom AI & Gen AI Development
End-to-end development of bespoke AI models tailored to your domain. Architecture, training, fine-tuning, and deployment — no shortcuts.
In-house model training
Agentic AI & Workflow Automation
Multi-agent systems that reason and automate complex workflows. RAG pipelines to autonomous AI agents integrated into your existing stack.
Agentic AI
🚀
AI Product Building for Startups
We partner with early-stage and funded startups as their embedded tech team — scalable AI products built with your runway and constraints in mind.
Startup-first
🏷️
Data Training & Annotation
In-house data operations — labeling, annotation, structuring unstructured data, and building high-quality training datasets that make models work.
In-house capability
🤝
Tech Partnership & Co-build
We become an integral part of your team — advising on architecture, solving hard tech challenges, building with the ownership of founders, not vendors.
Embedded team model
🎓
Adaptive Learning with Gen AI
Building next-gen adaptive learning platforms powered by Gen AI — personalising education at scale. Active development with a stealth EdTech startup.
Active project
Proof of Work

Built. Shipped.
Trusted by Enterprise.

9+ years of AI/Gen AI — our own models, in-house data pipelines, and production deployments at India's most recognised organisations.

Live AI Pipeline — Fusion Hiring
Resume Upload
NLP Parsing
Embedding Model
Ranked Output
JD Input
Similarity Score
Bias Filter
Shortlist ✓
Flagship Product · AI/ML Applicant Tracking System
Fusion Hiring
Our AI/ML-powered ATS — built entirely in-house — uses custom recommendation engines, CV-to-JD matching, talent resurrection from historical hiring data, and a Google-like search across candidate databases. A live demonstration of our Gen AI capability at enterprise scale.
Visit fusionhiring.com →
90%
Match accuracy
85%
Junk CV filtered
95%
Non-Match accuracy
35%
Faster time-to-hire
Enterprise Client · Banking & Financial Services
Kotak Mahindra Bank
Deployed Fusion Hiring's AI engine within one of India's leading private sector banks, transforming high-volume recruitment with intelligent screening and data-driven shortlisting.
Kotak Mahindra Bank BFSI · Enterprise
Enterprise Client · Infrastructure & Quality
Tata Projects Group
Delivering AI-driven talent operations across multiple entities within the Tata Projects ecosystem — intelligent, scalable recruitment within India's most trusted conglomerates.
TQCert TPLSPL Tata Projects
🎓

Stealth EdTech Startup — Adaptive Learning Platform

Currently co-building a Gen AI-powered adaptive learning system with an early-stage EdTech startup — personalising learning journeys at scale. Details under NDA.

How We Work

Not a Vendor.
A Partner.

01

Exploratory Call — No Pitch

We start by deeply understanding your problem, not selling you a solution. One honest conversation to see if we're the right fit.

02

Architecture & Feasibility

We map out the right AI approach — custom model, fine-tuned LLM, agentic pipeline — with a clear cost-to-value perspective.

03

Embedded Build Cycle

Our engineers integrate with your team. Regular syncs, shared context, and genuine ownership of outcomes — not just deliverables.

04

Scale Without Surprises

We design for scale from day one — infrastructure, model retraining pipelines, and monitoring baked in from the start.

🧠
🚀
🤝
"
We understand the challenges of a startup — the journey is not easy. We build solutions as our own, keeping financial constraints in mind while delivering scalable impact at optimum cost.
— Sonia Uppal & Dwaipayan Som, Co-Founders
In-house data annotation & model training — zero third-party dependency
9+ years of focused AI/Gen AI experience, not a recent pivot
Co-founders who've built and sold a startup — we think like you
Enterprise-validated across BFSI and Infrastructure sectors
Scalable cost model designed for startups through to large enterprises
Who We Serve

An Idea is Enough
to Start.

💡
Idea-Stage Founders
You have a conviction, maybe a wireframe. We help validate, architect, and build your first AI-native product without burning your runway.
📈
Funded Startups
You've raised and need to move fast. We act as your embedded AI engineering team — deep ownership without the overhead of hiring in-house.
🏢
Enterprises & Corporates
Infuse AI into workflows, automate decisions, build custom models on proprietary data. Validated at Kotak Mahindra Bank and Tata Projects.
🔄
Automation Seekers
Repetitive workflows, slow data processing, manual document handling — we build agentic AI systems that remove friction permanently.
Client Voices

What Our Clients
Actually Say

Thanks to the Fusion Hiring engine, we have been able to filter 85% junk. Additionally, decision making can be attained with few sample hired resumes.
SD
Sumit Kumar Das
Chief Manager — Human Resources
Kotak Mahindra Bank
Making the data from unstructured to structured helped the recruiter find the key points they are looking for. The tool with intelligence helped reduce workload and narrow down search to right and relevant profiles.
LC
Lalit Chacko
Director
Cognicx IT Solutions
The People

Founders Who've
Built Before.

SU
Sonia Uppal
Co-Founder · Strategy & Client Partnerships
Veteran of Fortune 500 companies turned founder. Sonia bridges deep enterprise relationships with startup agility — driving strategy, sales, and the client partnerships that define long-term success.
Enterprise Strategy Client Relations IT Leadership
DS
Dwaipayan Som
Co-Founder · AI/ML & Engineering
Serial entrepreneur — founded and sold a startup before Curioso. Dwaipayan lives at the intersection of AI/ML, performance engineering, and IoT. He doesn't just advise on AI — he architects and builds it.
AI / ML Gen AI Performance Eng. IoT
Let's Talk

One Call.
That's All It Takes to
Find Out If We're Your Team.

No pitch deck. No pressure. Just an honest conversation about your idea, your problem, and whether we're the right people to help you build it.

Headquarters
Bangalore, India · 560076
About Curioso

What Is Curioso
Technologies?

Curioso Technologies is an AI product studio headquartered in Bengaluru, India. Founded by Sonia Uppal and Dwaipayan Som, Curioso builds custom Generative AI products, agentic AI systems, and LLM-powered applications for funded startups and enterprise clients. The company has been building AI and machine learning solutions since 2018, accumulating over 9 years of focused AI engineering experience.

Curioso operates as an embedded technology partner — not a consulting firm or a reseller. The team designs, trains, and deploys custom AI models in-house, maintaining full ownership of data pipelines, model training, and production infrastructure. This model allows Curioso to serve both early-stage startups and large enterprises with the same depth of capability.

Curioso's flagship product is Fusion Hiring — an AI-powered applicant tracking system deployed at Kotak Mahindra Bank and Tata Projects subsidiaries. Fusion Hiring demonstrates Curioso's ability to build production-grade AI products at enterprise scale, filtering 85% of irrelevant CVs using NLP and semantic matching.

9+ yrs
AI / Gen AI focus
2 founders
Serial entrepreneurs
100 % happy clients
Enterprise validated
100%
In-house AI

How Curioso Compares

Capability Curioso Technologies IT Consultancy Freelance Dev
In-house model training ✓ Yes — custom models ✗ Uses third-party APIs ✗ Rarely available
Agentic AI systems ✓ Multi-agent pipelines ~ Depends on team ✗ Limited scope
Data annotation pipeline ✓ Fully in-house ✗ Outsourced ✗ Not included
Enterprise references ✓ Kotak · Tata Projects ~ Varies widely ✗ Typically none
Startup-friendly pricing ✓ Designed for startups ✗ Enterprise pricing ✓ Flexible rates
Production AI product (proof) ✓ Fusion Hiring (live) ~ Project portfolio ~ GitHub repos
FAQ

Questions About
Curioso Technologies

Everything you need to know about working with Curioso — answered clearly.

What does Curioso Technologies do?

Curioso Technologies is an AI product studio based in Bengaluru, India that designs, builds, and deploys custom AI products for startups and enterprises. Services include Custom AI and Generative AI product development, Agentic AI and workflow automation, AI product co-builds for funded startups, enterprise AI deployments, in-house data annotation and model training, and adaptive learning systems using Gen AI. Curioso acts as an embedded technology partner — not a vendor — taking co-founder-level ownership of every product it builds.

How is Curioso Technologies different from other AI development companies?

Three things set Curioso apart. First, in-house model training — Curioso builds and trains custom AI models from scratch using its own data annotation pipelines, without relying on third-party APIs. Second, enterprise proof — Curioso has deployed production AI at Kotak Mahindra Bank and Tata Projects, two of India's most demanding enterprise environments. Third, startup economics — Curioso's pricing and engagement model is designed to work within startup runway constraints, not just large IT budgets. The founders have previously built and sold a startup themselves, which shapes how they work with founders.

Who are Curioso Technologies' clients?

Curioso's enterprise clients include Kotak Mahindra Bank, one of India's leading private sector banks, and Tata Projects subsidiaries TQCert and TPLSPL. Curioso is also actively co-building an adaptive learning platform with a stealth EdTech startup. The company serves funded startups across India and globally, as well as non-IT enterprises looking to automate operations with custom AI — particularly in BFSI, infrastructure, manufacturing, logistics, and education.

What is Fusion Hiring and is it built by Curioso?

Yes. Fusion Hiring is Curioso Technologies' flagship AI product — an AI-powered applicant tracking system (ATS) built entirely in-house. It uses NLP and semantic matching to screen CVs against job descriptions, filtering 85% of irrelevant applications and delivering a ranked shortlist in minutes. It also includes a Talent Resurrection engine that resurfaces strong candidates from historical CV databases, and a JD Recommendation Engine that generates job descriptions using generative AI. Fusion Hiring is deployed at Kotak Mahindra Bank and Tata Projects. Learn more at fusionhiring.com.

Can Curioso build a custom AI product for an early-stage startup?

Yes — and this is one of Curioso's core strengths. Curioso specifically designs its engagement model to work with early-stage and funded startups, offering flexible cost structures that align with runway constraints. Co-founders Sonia Uppal and Dwaipayan Som have previously founded and exited a startup, so they understand the pressures founders face. Curioso acts as an embedded tech partner — joining your team, owning outcomes, and building with the urgency of founders rather than the pace of a consulting firm. An initial exploratory call is always no-pitch and no-pressure.

What is agentic AI and does Curioso build it?

Agentic AI refers to AI systems that can reason, plan, and take sequences of actions autonomously — rather than just responding to single prompts. Examples include multi-agent pipelines that automatically research, decide, and execute tasks across multiple systems, or RAG (Retrieval-Augmented Generation) systems that answer questions by pulling from live databases. Curioso builds bespoke agentic AI systems integrated into your existing tech stack, with monitoring and retraining pipelines included from the start.

Where is Curioso Technologies located?

Curioso Technologies is headquartered in Bengaluru, Karnataka, India (PIN 560076). The team works with clients across India and internationally — including the US, UK, and the Middle East. Co-founders Sonia Uppal and Dwaipayan Som can be reached at sonia.uppal@curiosobyte.com.

How do I get started with Curioso Technologies?

The best way to start is to book a free, no-pitch exploratory call with the Curioso team. The call is designed to understand your problem — not to sell you a solution. From there, Curioso will outline a feasibility and architecture proposal, explaining the right AI approach for your use case with a clear cost-to-value perspective. You can book directly via the calendar on this page, or email sonia.uppal@curiosobyte.com.

Still have questions?
📧 sonia.uppal@curiosobyte.com
📞 +91 966 337 2620