DataHulk.Let’s talk ↗
FORWARD DEPLOYED AI ENGINEERING

From client problem
to deployed
AI solution.

I work alongside enterprise teams to turn business requirements into governed AI solutions—with Claude, Databricks, Google ADK and a spec-driven engineering discipline.

Explore the work ↗Discuss your challenge →
Manoj Kumar
Manoj KumarForward Deployed AI Engineering Architect (FDEA)
DataHulk futuristic illustration bringing together cloud, data and AI technologies
THE DATAHULK UNIVERSE01 / ∞
Engineering insights.
Delivering impact.↗
CORE ENGINEERING
FOCUS
Claude & Claude Code
Databricks
Google ADK
Spec-Driven Development
01 / SPECIALIST EXPERTISE

Four technical disciplines.
One delivery mindset.

Forward Deployment Engineering connects client discovery, solution architecture, hands-on implementation and adoption. These are the capabilities I bring to that work.

✳AI-ASSISTED ENGINEERING

Claude & Claude Code

Design Claude-powered workflows and help teams adopt AI-assisted software engineering. Establish project context, reusable skills, MCP integrations, permission boundaries and review practices.

CLAUDE.mdSkills & MCPAmazon Bedrock
Discuss Claude enablement ↗
▱GOVERNED DATA & AI

Databricks engineering

Connect lakehouse modernization with agentic use cases. Design reliable data pipelines, Unity Catalog governance, RAG foundations and MLflow evaluation around business requirements.

Delta LakeUnity CatalogMosaic AI & MLflow
Explore lakehouse case studies ↗
⌘AGENT WORKFLOW DESIGN

Google ADK

Structure agent workflows around specialist responsibilities, tool contracts and evidence. Explore coordinator-led investigations, state management and human approval boundaries.

Agent orchestrationTools & stateEvaluation
Explore the ADK case study ↗
↳REQUIREMENTS INTO ENGINEERING

Spec-Driven Development

Translate business intent into reviewable specifications, acceptance criteria and implementation tasks. Use OpenSpec with AI coding workflows to keep delivery aligned with agreed behavior.

OpenSpecAcceptance criteriaTraceability
Explore the delivery approach ↗
02 / SAMPLE SOLUTION CASE STUDIES

Engineering patterns.
Evidence in the code.

Seven personal R&D case studies exploring agent orchestration, Databricks-oriented lakehouse patterns and governed actions. These are sample solutions, with simulated components clearly identified.

Explore GitHub ↗

Showing 7 projects

03 / FORWARD DEPLOYMENT ENGINEERING

Embedded with your team.
Accountable to the use case.

My delivery approach connects the client’s operational problem to a working system. Specifications guide the build; evaluations and user feedback guide the next iteration.

01 / DISCOVER

Understand the workflow

Work with stakeholders to map users, integrations, data access, risks and success measures.

02 / SPECIFY

Agree on behavior

Define requirements, architecture decisions, acceptance criteria and a prioritized task plan before implementation.

03 / ENGINEER

Build in context

Implement the chosen Claude, Databricks or ADK components with explicit tool contracts and integration tests.

04 / VALIDATE

Test the boundaries

Evaluate task quality, permissions, failure handling, latency and cost against the agreed specification.

05 / DEPLOY & ITERATE

Close the feedback loop

Plan rollout, observability and runbooks. Use operational feedback to refine the specification and enable the team.

SPEC-DRIVEN DEVELOPMENT × LOOP ENGINEERING

Agree. Build. Evaluate.
Learn. Repeat.

A specification is a shared engineering contract. My approach links requirements to tasks, implementation evidence and acceptance checks, then feeds what we learn back into the next change.

Business intentReviewable specsTest evidenceHuman review
01

OpenSpec

Structure specifications and changes for AI-assisted development.

↗02

Claude project context

Carry engineering instructions into the coding workflow.

↗03

Google ADK

Explore the agent framework and its documented patterns.

↗Technical references · Framework choice follows the use case.
Manoj Kumar, founder of DataHulkBENGALURU, INDIA · GLOBAL DELIVERY
04 / MEET YOUR FDE PARTNER

Client context.
Engineering depth.

I’m Manoj Kumar, founder of DataHulk and a Forward Deployed AI Engineering Architect (FDEA).

With 18+ years in enterprise AI, cloud and data transformation, I bridge client problems, architecture, hands-on engineering and production deployment. My specializations include FDE, Loop Engineering, agentic AI, RAG, MCP and AI-assisted software engineering.

I combine enterprise advisory with instructor-led capability building across Claude, Databricks and multi-cloud platforms. Engagements can include architecture reviews, focused proofs of concept, role-based workshops and implementation mentoring.

FDE & Loop Engineering SpecialistEnterprise AI Transformation ArchitectPrincipal AI & Cloud Architect
Connect on LinkedIn ↗Watch on YouTube ↗
18+Years of experience
25+Enterprise transformation programmes
10,000+Advisory & enablement hours
1,000+Professionals upskilled
EXPERIENCE

Architecture meets enablement.

  • UNext For EnterpriseDatabricks Architect & Senior SME — Cloud, Data & AI Consulting
  • NIIT LimitedEnterprise learning, cloud & capability transformation
  • K2View & Informatica UniversityGlobal and APAC technical enablement
  • Independent ConsultingEnterprise AI, data and cloud advisory & enablement
SELECTED CREDENTIALS

Specialist technical foundation.

  • Claude Certified Architect – Foundations
  • Claude Certified Architect – ProfessionalIn progress
  • Databricks Certified Generative AI Engineer Associate
  • Databricks Certified Data Engineer Associate
  • Microsoft Certified Trainer (MCT)

Profile and career figures supplied in Manoj’s résumé. Case-study prototypes are presented separately from professional engagements.

LET’S BUILD WHAT’S NEXT

Bring your use case.
Let’s engineer the path.

Scope an FDE engagement, build with Claude or Google ADK,
modernize Databricks, or introduce Spec-Driven Development to your team.

Start a conversation ↗Connect on WhatsApp ↗