Translation Model Benchmark for Multilingual Video Transcripts
A multilingual benchmark comparing Google Translate, DeepL, and Llama Maverick 4 on noisy video transcript data across 15 languages.
bakalis.io
Nikolas Bakalis
Senior Software Engineer, Applied AI & Backend
I build production AI, ML, and backend systems end-to-end, from early product development through scaled production operations, covering problem framing, evaluation, deployment, monitoring, and business impact.
- Current role Senior Software Engineer
Applied AI and backend systems at Channel Factory in New York.
- Experience 5+ years
Shipping production systems across AI, ML engineering, backend, and full-stack work.
- Recognition Tech Star
Named Product/Tech Star of the Year 2025-2026.
- Education M.S. AI & Big Data
Master's in Artificial Intelligence and Big Data, plus an integrated B.Eng. and M.Eng. in Computer Engineering.
Technical work
Reports, benchmarks, and implementation notes
RDS vs StarRocks 20M Serving and Aggregation Benchmark
A 20 million row benchmark comparing RDS/Postgres serving tables with StarRocks OLAP tables and async materialized views.
Materialized View API Serving Benchmark
An API-level comparison of denormalized RDS tables and StarRocks async materialized views across 100k, 1m, and 10m row scales.
IAB 3.0 Content Classifier Training Report
A training report for an in-house hierarchical IAB 3.0 content classifier built to replace external classification dependencies.
API Framework Benchmark for Data-Intensive Services
A benchmark of API frameworks and datastore access patterns for data-intensive services spanning PostgreSQL, StarRocks, and OpenSearch.
IAB 3.0 Classification and Language Detection Pipeline Proposal
A project proposal for replacing external classification APIs with in-house IAB 3.0 classification and language detection pipelines at media-corpus scale.
Adversarial Robustness in Medical Image Classification
Master's thesis work evaluating adversarial attacks and defensive training strategies for cervical cancer screening image classification models.
What I do
Systems I build
My work usually sits at the intersection of applied AI, backend services, and data infrastructure: defining the problem, building the system, measuring behavior, and operating it after launch.
Production AI Systems
I turn models into evaluated, monitored product workflows.
- Frame use cases
- Build RAG flows
- Evaluate outputs
- Monitor behavior
Backend Product Systems
I design APIs, workers, and service contracts for real product use.
- Design APIs
- Build workers
- Tune latency
- Harden releases
Data & Serving Platforms
I build the storage, analytics, and serving paths behind AI decisions.
- Model data flows
- Build lakehouse paths
- Serve analytics
- Scale search
End-to-end delivery
- Problem framing
- Evaluation
- Observability
- Production ownership
New York, NY
Professional profile
My work sits between applied AI, backend engineering, and product-facing delivery. At Channel Factory, I build applied AI, automation, and data systems across early product development and scaled production workflows, working with datasets spanning tens of millions of YouTube channels and billions of videos. Before that I shipped ML and full-stack systems for logistics and life-sciences workflows. This site is the deeper version of my resume: it shows how I evaluate tradeoffs, measure systems, and turn ambiguous technical problems into working software.
Channel Factory
Nov 2024 - PresentSenior Software Engineer, Applied AI & Backend
- Builds production AI, data, and backend systems for advertising workflows across tens of millions of YouTube channels and billions of videos.
- Delivers early-stage MVPs and scaled production systems with measurable revenue and operational impact.
- Works with foundation models, RAG, MCP, data lakehouse architecture, and service contracts.
ACERTUS
May 2022 - Nov 2024Software Engineer, ML Engineering & Full Stack
- Built production ML for ETA, delay, and fraud prediction in logistics workflows.
- Owned cloud deployments with Python, AWS, Docker, and production monitoring.
Vivpro
Jun 2021 - May 2022Full Stack Engineer, ML Engineering
- Built ML and full-stack systems for regulatory and clinical-trial intelligence workflows.
- Worked on natural-language interfaces before commercial LLM tooling became mainstream.
Education
- M.S. Artificial Intelligence and Big Data, Anglia Ruskin University
- Integrated B.Eng. and M.Eng. Computer Engineering, University of Patras
- Thesis work across adversarial attacks, image classification, and database tooling
