Melvin Akoth
Nairobi
Summary
I am a full-stack developer focused on building and deploying AI-integrated applications, automation, and robotics-adjacent systems with a focus on continuous knowledge expansion in the field of artificial intelligence.
Overview
4
4
years of professional experience
Work History
Co-Founder, Full-Stack Developer
MIRALITY
Dublin
09.2025 - Current
- Developed a learning-based robotic manipulation pipeline on the LeRobot SO-101, integrating teleoperation workflows with real-world robotic execution.
- Conducted exploratory research on NVIDIA Isaac Sim simulation as part of sim-to-real experimentation, assessing its suitability for teleoperation, synthetic data generation, and policy validation alongside MuJoCo.
- Contributed to the delivery of a customer-facing robotics web platform.
- Designed and implemented real-time teleoperation control interfaces using WebRTC, enabling low-latency bidirectional video and control streams for remote robotic arm operation.
Manager
INGUBU KENYA
Nairobi
09.2025 - Current
- Built the Kenya GTM motion for Fintech, Insurance, and NGO segments; defined ICPs for new clients.
- Drove customer discovery efforts through cold outreach and direct engagement with prospective clients, working with executive and functional stakeholders to surface pain points and shape prioritized use cases.
Ambassador
NVIDIA
EMEA
08.2022 - Current
- Organized tech webinars and events targeting university students to showcase new technologies and innovations from NVIDIA with a focus on GPU technology and robotics.
Full-Stack Developer
INGUBU
Madrid
04.2022 - Current
- Designed and prototyped agentic automation workflows using Sim.ai, evaluating orchestration patterns, tool-calling strategies, and reliability under real-world operational constraints.
- Integrated explainable AI (XAI) into production models using SHAP and LIME, delivering audit-ready feature attributions that increased stakeholder trust and accelerated model approval and deployment.
- Orchestrated federated learning pipelines across distributed clients with secure aggregation and privacy controls, maintaining model performance while meeting data-residency and compliance requirements.
- Engineered graph machine learning systems by designing and deploying knowledge graphs and graph-based algorithms, training GNN models (Node2Vec, Graph Attention Networks) to achieve measurable recommendation signal improvements over non-graph baselines.
- Developed a hiring automation API to extract, normalize, and score candidate data across multiple sources, reducing manual recruiter review and achieving a ~30% reduction in time-to-decision.
- Implemented an image forensics detection pipeline using Error Level Analysis (ELA) and pixel-level heuristics to identify manipulated media, supporting automated trust and content-integrity checks in AI workflows.
- Built a Generative BI (GenBI) AI agent enabling natural-language analytics (text-to-SQL, auto-generated charts, spreadsheets, and dashboards), empowering non-technical teams with self-serve business intelligence.
Education
Bachelor of Science - Mathematics and Computer Science
Technical University Of Mombasa
Mombasa, Kenya06.2023
Skills
Core AI & ML
Machine Learning (supervised, unsupervised)
Natural Language Processing (NLP)
Computer Vision & Multimodal Models
Graph Machine Learning (GNNs, knowledge graphs)
Advanced & Specialized
Federated Learning
Reinforcement Learning (applied)
Vision-Language-Action Models (VLAs)
Explainable AI (XAI)
Robotics & Embodied AI
LeRobot SO-101 robotic arms (Hugging Face)
Vision-to-Action pipelines
Simulation-to-real workflows
Systems & Automation
Agentic automation systems
AI workflow orchestration
Data pipelines & model integration
Timeline
Co-Founder, Full-Stack Developer
MIRALITY
09.2025 - Current
Manager
INGUBU KENYA
09.2025 - Current
Ambassador
NVIDIA
08.2022 - Current
Full-Stack Developer
INGUBU
04.2022 - Current
Bachelor of Science - Mathematics and Computer Science
Technical University Of Mombasa