Video summary

6 Hackathon Wins to ML Engineer at Kapture CX | Sankalp's Self-Built Success Story

Main summary

Key takeaways

Educational

Main ideas / lessons conveyed

  • Sankalp’s progression from curiosity to competitive success

    • Began programming in 8th grade, sparked by exposure to a Linux terminal on his sister’s laptop.
    • Developed interest through YouTube and started exploring cybersecurity/ethical hacking, especially CTFs (Capture the Flag).
    • Built breadth by exploring multiple areas: cybersecurity → web/front-end → React → back-end → ML.

  • Building a programming community at college

    • In early college years, there was no strong programming culture/coding club.
    • Sankalp formed a college-wide programming club, recruited teammates, and encouraged members toward:
      • hackathons
      • ideathons
      • competitions
    • The community scaled up, leading to inter-college tech events and multiple championships.

  • Hackathons as the learning engine (and how winning was achieved)

    • Participated in many hackathons (claiming ~6 hackathon wins plus additional ideathons/other wins).
    • Early challenges:
      • applications were often rejected
      • ideas were sometimes “already made”
    • Turning point:
      • later hackathons required pitching an idea, which improved his ability to learn from the patterns of winning projects
      • realized early work was too generic and needed specialization with clearer problem statements
    • First major win (college hackathon):
      • Theme: healthcare
      • Problem: build an automated system for patients to describe symptoms in Kannada/Hindi before consultation (supporting faster diagnosis via STT/TTS)
      • Mentioned technical approach:
        • training TTS text-to-speech
        • applying RAG (retrieval-augmented generation) to fetch relevant info from patient/past data for the doctor
        • creating a doctor-like “chatbot” interface and searching using EHR (electronic health record)
      • Result: won first prize, boosting motivation and confidence.

  • Industry gap: projects must be scalable and production-ready

    • During placements/interviews, he realized hackathon projects were often not scalable/production-ready.
    • He shared a “reality hit”:
      • an app that works locally can crash when scaled to hundreds/thousands of users.
    • He responded by taking work that forced real engineering standards.

  • Relevant real-world engineering experience

    • Worked with/through an NGO, then got a freelance project with Indian Railways.
    • Outcomes/learning (as described):
      • exposure to switches
      • writing CI/CD pipelines
      • production-grade coding practices
      • building an application used by ~500 users
    • This contrasted with interview-stage project expectations and helped him improve.

  • How he prepared for placements/internships

    • DSA practice
      • referenced standard interview question sets and contest practice
      • solved ~150–200 LeetCode problems and participated in contests
    • Networking + referrals
      • attended tech meetups every few months
      • got referrals through communities (his first interview reportedly came from a referral)
    • Job-search strategy
      • started on LinkedIn, then shifted to Wellfound (described as better for conversions)
      • stopped the “apply to everything quickly” approach
      • tailored resume per specific JD:
        • applied to fewer postings per day
        • adjusted resume/projects to match each role
    • Project presentation tip
      • recruiter advice: make projects live with clickable links
      • problem seen in others: projects only existed as GitHub repos

  • Capture CX internship: how the opportunity happened + interview details

    • Heard through the jobs team email forwarded via his college.
    • Applied via AQ Jobs, took a test (scored about 80%).
    • Interview emphasis
      • ML/DL fundamentals and applied topics
      • mentioned areas: recall classification, decision trees, random forest, chatbots, LLMs, transformers
    • Technical interviews
      • one with a manager (~1.5 hours, ML/DL-heavy)
      • one with an engineering manager (~40 minutes), focused on work and research interest
    • Current focus at Capture CX:
      • research & development
      • exploring emerging AI technologies from recent papers
      • improving existing systems by implementing new ideas
    • He states he’s working remotely.

  • Feedback on AI-based testing/interview experience

    • Liked the overall process, but noted a limitation:
      • AI tests/interviews sometimes lacked complete context and didn’t handle edge cases correctly
      • friends experienced issues where the AI claimed they hadn’t solved cases even after they did—suggesting missing test-case context.

  • Advice to job-prep ecosystem / companies like AQ Jobs

    • Suggestions focused on:
      • better visibility/reach to students (leveraging college placement departments and communities)
      • helping students move beyond outdated “YouTube clone” projects
    • Recommended student outcomes:
      • projects based on real domain problems
      • deployable, scalable, secure implementations (cloud, Docker, scalability/security—not just basic demos)
      • learning pathways that ensure projects aren’t ignored by recruiters

  • General success principle highlighted

    • Repeated theme: curiosity + consistent building + step-by-step improvement over time (not instant breakthroughs).

Methodology / instruction lists (as conveyed)

A) How to grow from interest to employable skills (implicit steps)

  • Start with curiosity-driven exploration (e.g., Linux → cybersecurity → web → ML).
  • Use competitions/CTFs/hackathons to convert learning into practice.
  • Create or join a team/community to increase participation and feedback.
  • Iterate from broad ideas to specialized, well-scoped problem statements.
  • Validate learning against real engineering standards (scalability, production readiness).

B) Hackathon strategy improvements learned by Sankalp

  • Apply consistently even when early rejections happen.
  • When pitching becomes required, learn from the structure of winning solutions:
    • how winning teams present
    • how they define problem statements
    • how they specialize instead of targeting broad/generic needs
  • Prefer specific impact projects over generic “works for everyone” apps.

C) Internship/placement preparation workflow described

  • Networking
    • attend tech meetups
    • build relationships to get referrals
  • DSA preparation
    • solve a target number of problems (referenced ~150+)
    • practice contest-style questions weekly
    • expect role-related topics (trees/arrays/stacks/queues/linked lists)
    • focus on domain-relevant questions rather than only extreme graph theory
  • Resume/job application strategy
    • stop mass-applying quickly
    • tailor resume to each job description
    • apply to fewer roles per day with more customization
  • Project visibility
    • deploy projects and share live links (clickable demos) instead of only GitHub repos
  • Feedback loop
    • after rejections, identify gaps (from feedback or self-review) and update projects/DSA accordingly

D) What makes a project recruiter-friendly (explicit advice)

  • Ensure the project stands out via one or more of:
    • a real-world problem you identified
    • thoughtful architecture/complexity
    • scalability or deployment maturity
    • deployable demo (live link)
  • Make projects easy to evaluate:
    • include a clickable link that proves the claimed functionality actually works

Speakers / sources featured

  • Sankalp (main interviewee; ML/internship candidate; placed as intern at Kapture CX / Capture CX)
  • Interviewer / host (AQ Jobs interviewer; appears as “Sir”/host asking questions)
  • Capture CX CTO (mentioned as someone who provides keen observations and is associated with interview/hiring)
  • Capture CX HR / Arsal (mentioned during the hiring process)
  • Manager / Engineering Manager at Capture CX (conducted technical rounds)
  • AQ Jobs team (platform used for applying/testing; involved in outreach to colleges and job processes)
  • Indian Railways (freelance project experience mentioned)
  • Manipal Hospitals (healthcare partner context for his winning hackathon)
  • Southern Railways (usage context for his application with ~500 users)
  • Referenced educational sources / educators:
    • Andrew Ng (Machine Learning Specialization, Stanford/Coursera)
    • CampusX (YouTube educator mentioned for teaching approach)

Original video