Hi, I'm

Bhaskar Jyoti Saikia

Agentic AI LLMs & MLOps Data Science

Head of AI Engineering, UK and Ireland @ Kyndryl

Building agentic AI systems for regulated enterprise clients.

Python
TensorFlow
GCP / AWS
FastAPI
Spark
Agentic AI

About me Resume LinkedIn

drag me

Hello.

My fascination with AI started with Iron Man. Watching Tony Stark build J.A.R.V.I.S. is what set the direction. I've spent my career turning that curiosity into production systems.

Certifications

Continuous learning, formally backed.

Oracle Certified Java Professional
Oracle Cloud Architect Associate
Google ML Crash Course
ML by Andrew Ng
ML A-Z (Udemy)

The Count

A career measured in coffee, code and shipped systems.

0 Cups of coffee
0 Lines of code
0 Projects
0 Pizzas eaten

About

Hands-on engineering leader. I write the code, own the architecture, and translate the result into commercial outcomes.

As Head of AI Engineering for UK and Ireland at Kyndryl, I lead the development of agentic systems from scratch. I design, build and deploy end-to-end data and ML pipelines that power intelligent automation for enterprise clients across banking, defence and the public sector.

My stack spans Python, TensorFlow, scikit-learn, FastAPI, Flask, Scala and Spark for production-ready ML and LLM applications. I work daily across GCP and AWS for data integrity, security and access at scale.

I hold an MSc in Data Science from Queen Mary University of London. I focus on applying LLMs and agentic AI to commercially demanding problems with measurable business impact.

Core skills

Agentic AI / LLMsLangGraph, LangChain, OpenAI Agents, RAG, fine-tuning, MCP
Data SciencePyTorch, TensorFlow, scikit-learn, knowledge graphs
MLOpsKubernetes, Docker, Airflow, Vertex AI, SageMaker
CloudGCP, AWS, GKE, Cloud Run, Lambda
Data EngineeringKafka, Pub/Sub, ETL, streaming analytics
GovernancePolicy-as-Code (OPA), HITL, audit, Zero Trust

Career

A decade of senior engineering across capital markets, banking, defence and global media.

Kyndryl

Full-time · Oct 2025 - Present · 9 mos

London, United Kingdom
1. Head of AI Engineering, UK and Ireland
  
  Apr 2026 - Present · 3 mos
  - Own the architecture and delivery of a proprietary Agentic AI framework now in production with Tier-1 UKI clients.
  - Designed the orchestration layer managing state, memory and tool use across multi-agent systems, with the eval harness and observability that make non-deterministic agents safe to ship.
  - Built the Model Context Protocol (MCP) integration layer and the combined MCP-plus-OPA pipeline that unblocks regulated client engagements.
  - Moved engagements from PoC to mission-critical production, cutting operational overhead by up to 80% in targeted processes.
  - Hired, mentored and run the UKI AI engineering team. Code reviews, architectural decisions and production debugging stay part of my week.
2. Director, Data Science
  
  Oct 2025 - Present · 9 mos
  - Lead strategic data science initiatives delivering measurable business impact through advanced AI technologies.
  - Set the engineering bar and lifecycle standards for production data science across UKI accounts.
NBCUniversal

Manager, Machine Learning Engineering · Feb 2024 to Sep 2025
- First ML hire in Europe. Stood up the team, the stack, the eval discipline and the delivery practice from scratch.
- Drove the adoption of generative AI and agentic systems across editorial, product and platform teams; pushed work from pilot into production rather than letting it sit in PoC.
- Built dynamic multi-step reasoning workflows with LangGraph and LangChain on top of RAG pipelines for live media catalogues.
- Productionised the Top 10 For You personalisation models powering Peacock, Sky and NowTV.
- Ran cross-functional, geographically distributed teams of ML engineers, data scientists and platform engineers.
Accenture

Lead Machine Learning Engineer · May 2022 to Feb 2024

Clients: Lloyds Banking Group, Nationwide, Ministry of Defence UK
- Led delivery of AI initiatives in financial services and defence, with generative models, real-time insights and knowledge-graph integration for regulated, insight-driven applications.
- Deployed fine-tuned Llama for enterprise data interrogation using RAG and rigorous prompt engineering, with evaluations gating each release.
- Built end-to-end MLOps on AWS, GCP and Databricks with Kubeflow and Kubernetes.
- Published a widely-read MLOps and LLMOps whitepaper.
Quantexa

Senior Machine Learning Engineer · Sep 2019 to May 2022
- Early ML engineering hire at one of the UK's leading capital-markets graph analytics companies. Helped lay the foundations of the ML engineering function.
- Built and operated ML models and data pipelines for fraud detection, AML and financial-crime intelligence on top of knowledge graphs used by Tier-1 banks.
- Shipped production-grade ETL and ML services on Kubernetes and Docker with full CI/CD discipline.
Queen Mary University of London

MSc Data Science · 2018 to 2019
- Statistical data modelling, ML for clustering and classification, big data processing.
- Notable project: predictive trading-signal model for AAPL and GOOGL using an LSTM with sentiment overlay.
- Real-time face detection on a drone with a TensorFlow CNN deployed on AWS and GCP.
Atos (client: Deutsche Bank) and Edelweiss Financial Services

Software Developer · Jun 2015 to Aug 2018
- Built backend in Java and frontend in JavaScript for an investment-banking platform; designed schemas and SQL objects for fraud detection workflows.
- At Edelweiss (capital-markets broker-dealer): built fraudulent-transaction and credit-card abuse detection models over trading and settlement data.
- Built RESTful APIs and microservices for cross-system data exchange.
Sharda University, Greater Noida, India

BEng Computer Science · 2011 to 2015