Saroj Bartaula Applied AI and Product Engineering

Resume

A concise, application-ready record of my experience, skills, and selected work.

Applied AI · Backend Systems · Product Engineering

Saroj Bartaula

Applied AI and Product Engineer

Spain sarojbartaula152@gmail.com sarojbartaula.com github.com/sbartaula linkedin.com/in/man-on-mission

Applied AI and product engineer who turns ambiguous requirements into working software. Strong in Python backend systems, LLM and agent applications, model training and adaptation, API integrations, and applied computer vision—from architecture and experimentation through deployment, reliability, and iteration.

Download PDF One-page A4 · formatted for recruiter review and ATS parsing.

Technical Skills

AI and models: LLM applications, agents, RAG, tool calling, model training, dataset preparation, tokenization, Hugging Face, PEFT/LoRA, evaluation, guardrails

Backend: Python, FastAPI, Django/DRF, REST APIs, Pydantic, PostgreSQL, webhooks, authentication, third-party integrations

ML and delivery: PyTorch, TensorFlow, OpenCV, MediaPipe, pose estimation, Docker, Linux, GitHub Actions, CI/CD, Firebase, React/Next.js, Android

Professional Experience

Founder & Applied AI / Product Engineer

Tenslam Vision

Sep 2025–Present · Spain

  • Shipped Tenslam Gym for Android, combining real-time pose estimation, movement logic, and user-facing feedback; used by 20+ users with 5,000+ exercise reps processed.
  • Build and test motion-intelligence workflows, AI/backend infrastructure, and product experiments, owning architecture, deployment, and iteration under latency, privacy, and reliability constraints.

Co-Founder & ML / Product Developer

TenSlam

Jun 2023–Mar 2025 · Spain

  • Built tennis-focused video-analysis prototypes with Python, OpenCV, object-detection tools, and Django REST APIs for movement tracking and technical feedback.
  • Ran product experiments and customer discovery around making video analysis useful to players and coaches without specialist hardware.

Earlier experience: AI / Product—Box99 / TheHeCapp (2022–2023); Entrepreneur in Residence—Demium (2022); independent web/data work (2020–2022); backend development—Eventti (2018–2020).

Selected AI Engineering Work

AI Agent Chat

Built a FastAPI system integrating Slack, GitHub, and LLMs for PR intake, CI monitoring, merge gating, and failure summaries, with webhook verification, replay protection, deduplication, and spend caps.

Model Training & Adaptation

Built and trained a 423K-parameter TinyGPT from scratch on custom text for 3,000 CPU steps; also tested Hugging Face Qwen and PEFT/LoRA adaptation workflows.

Flowtrack

Built an open-source, local-first activity tracker with a browser dashboard and AI analysis/chat; data stays on the machine by default.

Recognition

3rd Place · Humans in the Loop ALIA Challenge · March 2026

Education

  • Computer Science — STC Higher Education, Malta · Partial completion (2016; no degree awarded)

Languages

English (fluent) · Spanish (conversational) · Nepali (native)