London, UK Right to Work in the UK 🎓 MRes Artificial Intelligence – University of Wolverhampton, 2026

Transfer Pricing AI & Automation Intern – COOPR AI / ORA Advisors

I’m an AI & Automation Engineer with a background in enterprise software engineering and a growing focus on applied AI, agentic systems and reliable LLM-powered workflows.

I enjoy building systems where AI models, Python, APIs, automation and human judgement work together to solve real operational problems — not just standalone AI demos.

My current work explores LLMs/SLMs, local models, structured outputs, evidence validation, confidence-based escalation and human-in-the-loop AI, while my previous software engineering experience gave me a strong foundation in enterprise automation, integrations, governance and production support.

COOPR AI / ORA Advisors | July 2026 – Present

Working on AI-assisted workflows for transfer-pricing and finance-related processes, exploring how LLMs and smaller local models can be used safely within real operational environments.

Current areas of work include:

• Building Python-based LLM/SLM workflows
• Experimenting with local models using Ollama
• Generating and validating structured JSON outputs
• Designing confidence-based escalation and human-review paths
• Performing evidence and factual-consistency checks
• Exploring AI-assisted document and invoice-grounded workflows
• Combining probabilistic AI outputs with deterministic validation
• Improving traceability, explainability and auditability of AI workflows

AI Agents Multi-Agent Systems CAMEL Tool-Based Workflows Human-in-the-Loop

LLMs SLMs Ollama Structured Outputs Prompt Engineering Validation Evaluation

Python Flask REST APIs Requests JSON SQLite Pandas Git/GitHub

SBERT SentenceTransformers BM25 Semantic Search spaCy Information Retrieval

Power Apps Power Automate Power BI SharePoint SQL Server Microsoft 365

Human Review Audit Logging Evidence Checks Confidence Escalation Failure Handling RBAC

Python | Flask | REST APIs | SQLite | JSON/JSONL | JavaScript | Git

Built an end-to-end reconciliation system that monitors an asynchronous warehouse feed and determines how external lifecycle updates should affect internal system state.

• Detects newly available warehouse events using persistent revision checkpoints
• Calls external REST APIs and retrieves internal SQLite state
• Compares external and internal versions, statuses and lifecycle evidence
• Makes controlled decisions:

UPDATE_INTERNAL
KEEP_INTERNAL
NO_ACTION
HUMAN_REVIEW

• Prevents stale external data from overwriting newer internal records

• Routes ambiguous conflicts to a human reviewer instead of guessing

• Records detailed audit logs explaining:

  • what changed
  • why a decision was made
  • what action was taken
  • whether the action came from automation or human judgement

I also built a Flask dashboard for event simulation, state monitoring, decision history and human-review resolution.

This project reflects my interest in reliable agentic systems, state management, API integration, explainability and human-in-the-loop automation.

MRes Artificial Intelligence Dissertation – Completed 2026

Python | CAMEL AI | LLMs | Groq API | SBERT | BM25 | spaCy | PyMuPDF | Pandas

Designed an explainable postgraduate course recommendation system that analyses a candidate's CV and generates ranked course recommendations.

CV
 ↓
Document Extraction
 ↓
Structured Candidate Information
 ↓
┌─────────────────────────────┐
│ Skills Agent                │
│ Experience Agent            │
│ Education Agent             │
└─────────────────────────────┘
 ↓
Structured Candidate Profile
 ↓
BM25 + SBERT Retrieval
 ↓
Ranked Course Recommendations
 ↓
LLM-Based Explanations

• Built a CV-processing pipeline using PyMuPDF and spaCy

• Created structured candidate profiles from unstructured resume data

• Implemented BM25 lexical retrieval

• Implemented SBERT semantic retrieval

• Developed hybrid and section-weighted ranking strategies

• Used CAMEL multi-agent workflows for candidate analysis

• Integrated LLM-based explanation generation

• Evaluated recommendations using:

  • Precision@5
  • Recall@5
  • F1@5
  • Jaccard@5
  • nDCG@5
  • MRR

• Conducted controlled ablation experiments to understand the contribution of individual agents and retrieval components

My dissertation is now completed, and the project remains one of my main explorations into multi-agent AI, information retrieval and explainable decision-support systems.

Python | NLP | LLMs | Groq API | Scikit-learn | Structured JSON

Built an AI workflow for automatically classifying software issues.

• Developed a traditional TF-IDF + Logistic Regression baseline
• Compared baseline performance against an LLM-based approach
• Used structured JSON responses for predictable model outputs
• Added schema validation and failure handling
• Implemented human-review flags for uncertain cases
• Evaluated classification performance using standard ML metrics

2022 – 2025

Before moving deeper into AI engineering, I worked on enterprise application development and business-process automation using the Microsoft ecosystem.

I worked directly with business users, process owners and technical stakeholders to understand operational problems and turn them into production-ready solutions.

• Built enterprise applications using Power Apps
• Developed automated workflows using Power Automate Cloud & Desktop
• Integrated applications with SQL Server, SharePoint, Dataverse, Microsoft Teams and APIs
• Built multi-stage approval workflows, dynamic routing and SLA escalations
• Implemented validation, exception handling and audit trails
• Worked with Entra ID / Azure AD, RBAC and security groups
• Supported development, testing, UAT, releases and production environments
• Used environment variables, connection references and managed solutions
• Supported Git/Azure DevOps-based release and CI/CD practices
• Built Power BI reporting and analytics solutions
• Worked directly with stakeholders throughout the full solution lifecycle

This experience is a major part of how I approach AI today: AI needs to work within existing processes, systems, controls and human decision-making.

LLMs SLMs AI Agents CAMEL Ollama Groq API Structured Outputs Prompt Engineering Human-in-the-Loop AI Evaluation Output Validation Retrieval & Ranking

Python Flask REST APIs Requests JSON SQLite Pandas NumPy Scikit-learn

SBERT SentenceTransformers BM25 spaCy Semantic Search Cosine Similarity PyMuPDF

Power Apps Power Automate Power BI SharePoint Online Dataverse Microsoft 365 SQL Server Entra ID / Azure AD

Git GitHub Azure DevOps PowerShell API Integration Testing Logging Exception Handling Dependency Management CI/CD

I'm particularly interested in opportunities involving:

• AI Engineering
• Applied AI
• AI & Intelligent Automation
• Agentic AI / AI Agents
• LLM & SLM Applications
• Python Engineering
• API & Systems Integration
• Human-in-the-Loop AI
• Enterprise AI Automation
• Microsoft AI & Power Platform

I’m especially interested in teams building AI systems that need to be reliable, explainable, integrated with real business systems and useful to actual users.

University of Wolverhampton, UK | 2025 – 2026 | Completed

Focus areas:

Applied AI • AI Agents • Information Retrieval • NLP • Machine Learning • Recommender Systems • AI Evaluation

Dissertation: Hybrid IR–ML Framework for Personalised Course Recommendation via Resume Analysis Using CAMEL

Cochin University of Science and Technology, India | 2018 – 2022

First Class | CGPA: 8.24/10

Final Project: Intelligent Shopping Trolley – IoT & Embedded Systems

Agentic AI
LLM / SLM Application Engineering
Local Models with Ollama
AI Workflow Orchestration
Structured Model Outputs
Human-in-the-Loop Systems
AI Evaluation & Observability
Model Context Protocol (MCP)
Microsoft Foundry & Copilot Studio
Enterprise AI Integration

I’m currently exploring entry-level, graduate and internship opportunities across the UK in:

AI Engineer • Applied AI Engineer • AI Automation Engineer • Junior AI Engineer • Intelligent Automation Engineer • AI/Automation Analyst

I’m particularly interested in roles where I can combine my experience in AI, Python, APIs, automation and enterprise systems while continuing to grow as an AI engineer.

📍 London, UK Right to Work in the UK

LinkedIn https://www.linkedin.com/in/gopika-sushama

GitHub https://github.com/GOPIKA-SUSHAMA

Email gvndgpk@gmail.com