12 years · Principal AI Architect · shipping AI into production

Hi. I’m Shashank.
I make organizations ship with AI.

As Principal AI Architect at Taazaa I own the full arc of AI adoption—from infrastructure and model prompt architectures, the agent harnesses and custom hooks for Claude Code / Copilot / Cursor, the guardrails and AIBOM, and the sessions that turn every engineer into an AI-native one. I sit on AI governance, I write the playbooks, and I ship the internal platforms the org runs on.

Delivering enterprise solutions with AI is my jam. On nights and weekends I ship the platforms — 5 production LLM systems, including an 11-agent K8s root-cause swarm, a 35+ agent meeting-intelligence SaaS on EKS, and an AI governance platform with 394 test files.

12 years across UK banking, DACH HRMS, luxury e-commerce, and US field operations. I earn the room’s trust in the morning and ship into their VPC by evening. — Get busy building, or get busy shipping.

shash2577@gmail.comgithub.com/Shashank2577
scroll · what I do
01

What I do

Principal AI Architect @ Taazaa

Delivering enterprise solutions with AI is my jam.

I own AI strategy and adoption for the entire engineering organization. I bridge the gap between complex model architecture and practical governance, enablement, and the internal platforms that compound across every engagement. The personal projects below are what I build on nights and weekends.

01

AI Advisory & Strategy

Advise the org on where, how, and whether to use AI — from model selection to build-vs-buy to compliance posture.

+
  • ▸Author engineering playbooks, code guidelines, and architectural standards for AI-assisted delivery.
  • ▸Design standard AI architectures (RAG, agentic workflows, multi-modal ingestion)
  • ▸Advise on AI product incubation
  • ▸Authored the 65-project portfolio strategy adopted by Engineering Core
02

Prompt Engineering & Agentic SDLC

Delivering enterprise solutions with AI is my jam. I design the prompt architectures, agent harnesses, and custom hooks that make AI-assisted delivery reliable enough for production.

+
  • ▸Design prompt architectures and evaluation harnesses
  • ▸Build custom agent harnesses and hooks for Claude Code, Copilot, Cursor, Jules
  • ▸Roll out agentic development across every team with per-team measurement
  • ▸Track adoption %, cycle-time delta, and PR throughput
03

AI Governance & Guardrails

I sit on the AI governance side too — writing the policies, guardrails, and review processes that keep AI usage safe, compliant, and auditable.

+
  • ▸Author AI usage policies and guardrails
  • ▸Build AIBOM and audit-trail practices (EU AI Act aligned)
  • ▸Define OPA policies, KMS encryption, SSO/RBAC patterns for AI systems
  • ▸Review and approve AI features for production
04

Teaching & Enablement

I run the sessions that get engineers productive with AI — Cloud, GitHub Copilot, prompt engineering, building AI workflows. I've built internal sites and curricula around all of it.

+
  • ▸Deliver sessions on Cloud, GitHub Copilot, and prompt engineering
  • ▸Build internal teaching websites on Copilot, prompting, and AI workflows
  • ▸Author how-to-build-AI-workflows and custom-harness guides
  • ▸Mentor 12-member team on AI-native delivery
05

Internal AI Platforms

I build the internal platforms the org runs on — the AI-usage analytics, the SaaS accelerator, the meeting intelligence, the tools that compound across every engagement.

+
  • ▸Built multi-IDE AI-usage analytics (VS Code + IntelliJ + Cursor) — coverage LinearB/DX/Faros don't ship
  • ▸Architected 'cross-cutting' — 16-module Spring Boot SaaS accelerator, 264+ tests, 4 SDKs
  • ▸Deployed meeting intelligence (Spreix), drawing tools, transcription pipelines org-wide
  • ▸Cut SaaS engagement bootstrap from ~8 weeks to days
06

Incubation & Product

I incubate products from zero to production grade — the 5 crown jewels below all started as internal bets I championed, architected, and shipped.

+
  • ▸Incubated 5 products to production grade (see Projects below)
  • ▸Authored the portfolio strategy that prioritized them
  • ▸Own the build-vs-buy-vs-incubate decision for AI capabilities
  • ▸Run POCs, pilots, and technology evaluations for new bets
02

Skills

by layer

Languages

javagotypescriptpythonnodejs
Java (8–21)GoTypeScript / Node.jsPythonSQL

AI / LLM Engineering

anthropicopenaigeminilangchainopentelemetry
LLM application architectureMulti-agent orchestration (35+ agents in prod)MCP — 6 servers shippedRAG: pgvector, embeddings, HNSW, hybrid retrievalEvals & AI observabilityPrompt engineering & governance (AIBOM, EU AI Act)LangGraph, Vercel AI SDK, LiteLLMAgentic SDLC: Claude Code, Copilot, Cursor, Jules

Cloud & Platform

amazonwebserviceskubernetesdockerterraformhelmopentelemetryprometheusgrafana
AWS: EKS, Lambda, SQS/SNS, DynamoDB, Step Functions, S3, API Gateway, CloudWatch, X-RayKubernetes & Helm (charts published to GHCR)Docker, Terraform, GitHub ActionseBPF observability (Beyla, Tetragon)OTel, Prometheus, ClickHouse, VictoriaMetrics

Backend & Data

springgraphqlkafkaredispostgresqlelasticsearch
Spring Boot microservicesgRPC, REST, GraphQLKafka, ActiveMQ, BullMQ, RedisPostgreSQL, MySQL, ElasticsearchMulti-tenant SaaS architectureOAuth/SSO, RBAC, OPA, KMS

Frontend

reactnextjs
ReactNext.jsTailwind CSSFramer Motion
03

Projects

7 featured · 1163+ commits · 59 total

5 crown jewels (Tier 1) — shipped or near-shipped with defensible differentiation. The full 59-project archive lives on GitHub.

04

Experience

5 companies · 12 years

2024 – Present

Current

Principal Architect — AI

Taazaa Inc · Noida, India

Organization-wide AI transformation for a US software-product engineering firm

  • ▸Own AI strategy and adoption for the entire engineering organization — playbooks, standards, tooling, and measurement for AI-assisted delivery.
  • ▸Drove agentic-development rollout (Claude Code, GitHub Copilot, Cursor, autonomous agents like Jules) with per-team measurement of adoption, cycle-time delta, and PR throughput.
  • ▸Built the org's AI-usage analytics platform covering VS Code, IntelliJ AND Cursor — multi-IDE telemetry that commercial tools (LinearB, DX, Faros) don't offer — giving leadership per-team AI adoption, acceptance-rate, and impact data.
  • ▸Architected 'cross-cutting', a Spring Boot auto-configuration library: 16 production modules (multi-tenancy, RBAC, audit, pgmq queue, pgvector search, payments, LiteLLM AI gateway), 264+ tests, SDKs in React/Node/Python/Java — cuts SaaS engagement bootstrap from ~8 weeks to days.
  • ▸Authored the AI-SDLC playbook and engineering standards adopted as the org-wide delivery baseline. Advise on AI product incubation; authored the 65-project portfolio strategy adopted by the Engineering Core; incubated 5 products to production grade.
javaspringtypescriptkubernetes

07/2021 – 2024

Technical Specialist / Technical Architect

Coforge Ltd · Noida, India

Umantis (Haufe Group) — multi-tenant talent-management SaaS for the DACH enterprise market

  • ▸Architected and evolved a multi-tenant HRMS SaaS serving enterprise customers across Germany, Austria, Switzerland — Java 17 microservices on AWS EKS.
  • ▸Led a 12-member engineering team across design reviews, code review culture, POC evaluation, and mentoring.
  • ▸Owned platform architecture: gRPC + Eureka + circuit breakers, Kafka & ActiveMQ messaging, Elasticsearch search, MySQL/PostgreSQL with Flyway migrations, Redis caching, full AWS estate (Lambda, SQS/SNS, API Gateway, CloudWatch, X-Ray).
  • ▸Implemented security & GDPR-compliance practices for multi-tenant data isolation across DACH enterprise customers.
javaspringkuberneteskafkaredispostgresqlelasticsearch

08/2019 – 06/2021

Senior Software Engineer

AKQA · Gurgaon, India

E-commerce platforms for Richemont luxury brands: Panerai, Montblanc, Van Cleef & Arpels

  • ▸Client-facing engineer for 3 Richemont luxury maisons — owned architecture decisions and presented directly to brand-side stakeholders under exacting luxury-retail standards.
  • ▸Delivered the Panerai global e-commerce platform (16-month program): AEM (OSGi/Sling/JCR) + Java 8 microservices + AWS, with personalization, payment, and inventory integrations.
javaspring

05/2016 – 08/2019

Application Developer

IBM India · Noida, India

IBM US/Canada Parts Administration — ERP for field operations

  • ▸Built the real-time flagging system used daily by IBM field engineers across US/Canada parts operations — cut reporting from batch cycles to real-time, automated turnaround-time calculation, and surfaced cost-loss patterns.
javaspring

09/2014 – 05/2016

System Engineer

Tata Consultancy Services · India

Lloyds Bank (UK)

  • ▸Java developer on regulated UK retail-banking systems (19-month engagement) — first exposure to compliance-grade delivery.
java
05

Photography

beyond code

Photography taught me systems thinking. Light, composition, timing — architecture exists everywhere. The same eye that frames a shot frames a system. A growing collection, updated as the archive grows past a hundred frames.

156 of 156

Equipment

  • Cameras: Sony A7IV, Sony A7RIII
  • Lenses: Sony 24-70mm f/2.8 GM, Sony 85mm f/1.4 GM, Sony 35mm f/1.4 GM
  • Drone: DJI Mavic 3
  • Accessories: Peak Design Carbon Tripod, ProGrade Digital CFexpress Type A
06

Credentials & Contact

Certifications

  • AWS Certified Solutions Architect — Associate

Education

B.Tech — Mechanical Engineering

Hindustan College of Science and Technology · 2014

shash2577@gmail.com+91 9899304189github.com/Shashank2577linkedin.com/in/shashank-saxena-1a0463100observe.sovix.xyz
Noida, India · open to relocation (UAE / EU / SG) & remote

Visa: Employment visa (sponsorship) — UAE Golden Visa eligible via AI-specialist route. Languages: English (fluent), Hindi (native). Nationality: Indian.